PRIVILEGED AND CONFIDENTIAL — ATTORNEY WORK PRODUCT
INTERNAL MEMORANDUM
CALDERWOOD & HARKNESS LLP Healthcare & Life Sciences Practice Group 1401 K Street NW, Suite 900 Washington, D.C. 20005
From: Amy Whitfield, Lead Associate, Healthcare & Life Sciences Sean Morrissey, Working Associate, Healthcare & Life Sciences
To: Miriam Calloway, Responsible Partner, Healthcare & Life Sciences
Date: May 12, 2025
Re: Pre-Submission Strategy and Anticipated FDA Feedback — Solstice Lumina™ CGM System (510(k) Submission Planning)
Client: Solstice Biomedical Devices Ltd. / Solstice Biomedical Inc.
Engagement Ref. No.: 1037-00002
CONFIDENTIALITY NOTICE: This memorandum constitutes attorney work product prepared in anticipation of regulatory proceedings and is protected by the attorney-client privilege and the work product doctrine. It is intended solely for internal Calderwood & Harkness LLP use and should not be disclosed to any third party, including Pinnacle Regulatory Consulting Group LLC or Crestline IP Advisory LLC, without express authorization from the responsible partner. Any unauthorized disclosure, reproduction, or distribution of this document or its contents is strictly prohibited.
PRIVILEGED AND CONFIDENTIAL — ATTORNEY WORK PRODUCT
Right-click to update Table of Contents
PRIVILEGED AND CONFIDENTIAL — ATTORNEY WORK PRODUCT
I. Purpose and Executive Summary
This memorandum analyzes the likely positions of the U.S. Food and Drug Administration, Center for Devices and Radiological Health ("FDA" or "CDRH"), on each of the seven Pre-Submission ("Q-Sub" or "Pre-Sub") questions that Solstice Biomedical Inc. ("Solstice" or "SBI") intends to pose in its Pre-Sub meeting package. The Pre-Sub meeting package is targeted for submission by June 16, 2025, in advance of an anticipated meeting with the Division of Chemistry and Toxicology Devices within the Office of Health Technology 2 ("OHT2") on August 14, 2025. The Pre-Sub meeting request itself is targeted for submission by June 2, 2025, consistent with FDA's procedural requirements for scheduling.
The memorandum provides a risk-ranked assessment of each of the seven Pre-Sub questions, identifies the most probable FDA feedback on each question based on recent CDRH precedent in the continuous glucose monitoring ("CGM") regulatory space, and develops detailed contingency strategies for adverse outcomes. These adverse outcomes include, but are not limited to, potential rejection of the primary predicate device (the NovaSense GlucoTrack NX-500, 510(k) No. K211034), a determination by FDA that the device requires premarket approval ("PMA") rather than a 510(k) clearance, requests for additional or modified clinical endpoints beyond those currently contemplated in the LUMINA-US study protocol, demands for a full Investigational Device Exemption ("IDE") application in lieu of the non-significant risk ("NSR") exemption pathway, and questions regarding the appropriateness of integrated CGM ("iCGM") classification under the De Novo framework.
This analysis draws upon multiple sources of regulatory intelligence and legal precedent, including: publicly available 510(k) summaries and decision letters for devices cleared under product code QKQ; FDA guidance documents, with particular emphasis on the September 2023 final guidance titled "Integrated Continuous Glucose Monitoring Systems" and the September 2023 final guidance titled "Cybersecurity in Medical Devices: Quality System Considerations and Content of Premarket Submissions"; recent De Novo classification orders and 510(k) clearance decisions for CGM and related glucose monitoring devices; the March 2024 reclassification order for glucose test systems under 21 C.F.R. § 862.1355; and the professional experience of the engagement team. We note in particular the considerable value of Miriam Calloway's prior tenure as senior counsel at CDRH, which provides this engagement team with institutional insight into CDRH's internal review processes and policy perspectives that inform the assessments set forth herein.
Our overall strategic recommendation remains unchanged from the Regulatory Strategy Memorandum previously provided to the client on March 28, 2025: Solstice should pursue a traditional 510(k) submission covering both the Lumina C14 sensor and the Lumina R7 reader/transmitter as an integrated system, with a revised target of Q3 2026 for FDA clearance, consistent with the commercial launch timeline developed in coordination with Marcus Hale, VP of U.S. Operations. While certain risks identified in this memorandum warrant careful preparation and contingency planning, the overall regulatory posture of the Lumina CGM System is favorable, and the 510(k) pathway is both legally supportable and strategically sound.
PRIVILEGED AND CONFIDENTIAL — ATTORNEY WORK PRODUCT
II. Background
II.A. Device Overview
The Solstice Lumina™ CGM System (the "Lumina System" or the "Device") is a two-component integrated continuous glucose monitoring system designed for real-time glucose monitoring in adults with diabetes mellitus. The system comprises the following components:
Lumina C14 Sensor. The Lumina C14 is a subcutaneously implantable continuous glucose monitoring sensor designed for a 14-day wear duration. The C14 sensor incorporates a fourth-generation enzymatic electrochemical biosensor platform featuring Solstice's proprietary PtIr-4G platinum-iridium electrode coating, which Solstice's engineering team has developed to improve signal stability and reduce sensor drift over extended wear periods relative to conventional platinum electrodes. The C14 sensor communicates wirelessly via Bluetooth Low Energy ("BLE") 5.2 to the Lumina R7 reader/transmitter. The C14's measurement range spans 40–400 mg/dL of interstitial glucose concentration, with a warm-up period of 60 minutes or less following subcutaneous insertion.
Lumina R7 Reader/Transmitter. The Lumina R7 is a wearable, arm-worn reader and transmitter that serves as the primary user interface for the Lumina System. The R7 provides real-time glucose display, directional trend arrows indicating the rate and direction of glucose change, 8-hour trend graphs for retrospective review, and configurable hypoglycemia and hyperglycemia alerts with an alert latency of 5 minutes or less. The R7 communicates with the C14 sensor via BLE 5.2 and also communicates via BLE with the Solstice Lumina™ mobile application, available on both iOS and Android platforms. The arm-worn form factor of the R7 represents a departure from the handheld receiver design used in the primary predicate device (the NovaSense GlucoTrack NX-500) and is intended to improve user convenience and reduce the risk of missed alerts.
The Lumina System is intended for prescription use by persons aged 18 years and older with Type 1 or Type 2 diabetes mellitus for the management of their diabetes. The device is intended to replace fingerstick blood glucose testing for diabetes treatment decisions, including insulin dosing decisions, when used as directed by a healthcare professional.
The Lumina System incorporates the following software components:
• Lumina Firmware v3.4.2 — embedded firmware for the C14 sensor, controlling electrochemical signal acquisition, analog-to-digital conversion, and BLE data transmission.
• Lumina Reader OS v2.1.0 — operating system and application software for the R7 reader/transmitter, managing display rendering, alert generation, user interface, and BLE communication protocols.
• LuminaPredict Algorithm v2.1 — Solstice's proprietary glucose prediction and calibration algorithm, which processes raw electrochemical signals from the C14 sensor and generates calibrated glucose values, trend data, and predictive glucose alerts. The algorithm resides on both the R7 reader and the mobile application.
• Lumina Mobile App v1.0.3 — companion mobile application providing glucose data display, historical trend analysis, data export, and cloud synchronization functionality on iOS and Android platforms.
The Lumina CGM System received European Union Medical Device Regulation ("EU MDR") CE certification on September 12, 2024 (Certificate No. MCA-2024-07319), issued by Meridian Conformity Assessment GmbH, Notified Body 0483. The device has been commercially available in the European Union since September 2024, and early commercial performance data from the EU market has been generally consistent with the clinical performance data generated during the CE marking process.
Solstice Biomedical Inc. (State of Delaware, File No. 7834291), with its principal offices at 220 Westbrook Drive, Suite 400, Cambridge, Massachusetts 02142, is the intended 510(k) holder and the U.S. agent for Solstice Biomedical Devices Ltd. (Ireland), the parent entity and device manufacturer. All references in this memorandum to "Solstice" refer collectively to both entities unless otherwise specified.
II.B. Predicate Device Framework
Solstice's proposed predicate device framework for the 510(k) submission employs a split-predicate approach utilizing one primary predicate and one secondary (reference) predicate, supplemented by an additional reference device cited solely for purposes of establishing corporate pedigree and prior FDA interaction history. The predicate framework is as follows:
Primary Predicate: NovaSense GlucoTrack NX-500 CGM System. 510(k) No. K211034, cleared by FDA on June 18, 2022. Product code QKQ, Class II, classified under 21 C.F.R. § 862.1355 (System, Test, Glucose). The NX-500 is an enzymatic electrochemical continuous glucose monitoring system comprising a subcutaneously implanted sensor with a 10-day wear duration and a handheld BLE receiver/reader. The NX-500 employs a conventional platinum electrode and communicates via BLE 5.0. The NX-500 has an established safety and effectiveness profile and has served as a predicate for subsequent CGM clearances within the QKQ product code space.
Secondary Predicate (Reference Device): NovaSense GlucoTrack NX-700 CGM System. 510(k) No. K230489, cleared by FDA on March 7, 2024. Product code QKQ, Class II. The NX-700 represents an iterative advancement of the NX-500 platform, extending the sensor wear duration from 10 days to 14 days, upgrading BLE communication from 4.2 to 5.0, and implementing an updated proprietary algorithm. The NX-700 is cited as a secondary predicate primarily to bridge the sensor wear duration gap between the Lumina C14's 14-day wear and the NX-500's 10-day wear. The NX-700's 510(k) clearance establishes FDA's acceptance of a 14-day enzymatic electrochemical CGM sensor within the QKQ product code.
Additional Reference Device: Solstice GlucoStrip RX-200 Test Strip Reader. 510(k) No. K192847, cleared by FDA on November 14, 2019. Product code NBW, Class II. The RX-200 is a handheld blood glucose test strip reader previously manufactured and marketed by Solstice Biomedical Inc. in the United States. The RX-200 is cited solely to establish Solstice's corporate pedigree in the glucose monitoring device space and its prior regulatory interaction history with FDA's Division of Chemistry and Toxicology Devices. The RX-200 is not proposed as a predicate for purposes of demonstrating substantial equivalence.
The engagement team has identified five key technological differences between the Lumina System and the primary predicate K211034 (NovaSense GlucoTrack NX-500) that will require discussion in the substantial equivalence analysis:
1. Sensor Wear Duration: 14 days (Lumina C14) versus 10 days (NX-500). This difference is bridged by the secondary predicate K230489 (NX-700), which has an identical 14-day wear duration.
2. Electrode Material: Platinum-iridium alloy with proprietary PtIr-4G coating (Lumina C14) versus conventional platinum electrode (NX-500). This is the most significant technological difference and the one most likely to attract FDA scrutiny.
3. BLE Version: BLE 5.2 (Lumina System) versus BLE 5.0 (NX-500). This is an incremental wireless communication upgrade that is unlikely to raise substantial equivalence concerns but may trigger additional cybersecurity documentation expectations.
4. Reader Form Factor: Arm-worn integrated reader/transmitter (Lumina R7) versus handheld receiver (NX-500 reader). This represents a user-interface design change rather than a change in fundamental scientific technology.
5. Proprietary Algorithm: LuminaPredict v2.1 (Lumina System) versus the NovaSense proprietary algorithm (NX-500). Different algorithms for glucose calibration and prediction are a common feature of CGM 510(k) submissions and are typically addressed through performance data (MARD, error grid analysis) rather than as predicate-breaking differences.
II.C. Pre-Sub Questions Overview
The Pre-Sub meeting package will present seven categories of questions to FDA, each addressing a critical regulatory element of the planned 510(k) submission. These seven question categories, which are analyzed in detail in Section III below, are:
1. Predicate Acceptance and Substantial Equivalence Framework — seeking FDA's concurrence on the appropriateness of K211034 as the primary predicate and K230489 as the secondary predicate, and the adequacy of the proposed approach to addressing the five identified technological differences.
2. Clinical Study Design and Endpoints — seeking FDA's feedback on the proposed LUMINA-US clinical study design, including the study population, sample size, primary and secondary endpoints, and the acceptability of the MARD and consensus error grid performance targets.
3. IDE Exemption Applicability (NSR Determination) — seeking FDA's input on whether the LUMINA-US study qualifies for the IDE exemption as a non-significant risk study under 21 C.F.R. § 812.2(c)(3).
4. Software Documentation Requirements and Level of Concern Designation — seeking FDA's guidance on the appropriate software safety classification and the scope of software documentation expected for the four software components of the Lumina System.
5. Cybersecurity Premarket Submission Expectations — seeking FDA's confirmation that the planned cybersecurity documentation approach aligns with the September 2023 cybersecurity guidance and identifying any additional expectations specific to BLE 5.2 communication.
6. Biocompatibility Testing Scope Adequacy — seeking FDA's concurrence on the proposed biocompatibility testing battery for the C14 sensor (subcutaneous implant) and the R7 reader (skin-contacting external device), with particular attention to the PtIr-4G electrode material.
7. iCGM Classification Pathway Considerations — seeking FDA's input on whether the Lumina System may pursue iCGM classification through the 510(k) pathway and the specific performance thresholds and documentation requirements applicable to iCGM designation.
PRIVILEGED AND CONFIDENTIAL — ATTORNEY WORK PRODUCT
III. Analysis of Anticipated FDA Positions — Question by Question
III.A. Question 1: Predicate Acceptance and Substantial Equivalence Framework
III.A.1. Most Likely FDA Position
Based on our review of CDRH precedent in the CGM space, the current regulatory framework for product code QKQ devices, and the specific characteristics of the proposed predicate devices, we assess that FDA will likely accept K211034 (NovaSense GlucoTrack NX-500) as the primary predicate device for the Lumina CGM System. This assessment is grounded in several factors. First, K211034 and the Lumina System share the same product code (QKQ) and the same regulatory classification under 21 C.F.R. § 862.1355, Class II with special controls. Second, the intended use populations are identical — both devices are indicated for adults with Type 1 or Type 2 diabetes mellitus for continuous glucose monitoring. Third, and most importantly, both devices share the same fundamental technological characteristics: both are enzymatic electrochemical CGM sensors that measure interstitial glucose concentration via a subcutaneously implanted sensor element and transmit glucose data wirelessly via BLE to an external reader for display and alerting purposes.
The use of K230489 (NovaSense GlucoTrack NX-700) as a secondary or reference predicate is well-supported by CDRH precedent and serves an important bridging function. The NX-700's 14-day sensor wear duration is identical to the Lumina C14's wear duration, thereby addressing the most immediately apparent difference between the Lumina System and the primary predicate K211034 (which has a 10-day wear duration). The split-predicate approach — in which different features of the subject device are compared to different predicates to demonstrate substantial equivalence — is a recognized 510(k) strategy that CDRH has accepted in numerous prior CGM clearances and in other product code categories. We note that CDRH's own guidance on the 510(k) program acknowledges the acceptability of multiple predicate references and that the Agency has not taken the position that a single predicate must be identical in all respects to the subject device.
FDA's review will focus on whether the five identified technological differences between the Lumina System and the primary predicate K211034 raise different questions of safety or effectiveness. Our assessment of each difference is as follows:
Electrode Material (PtIr-4G vs. Platinum). This is the most likely point of FDA scrutiny and the difference most likely to generate detailed follow-up questions during the Pre-Sub meeting. The PtIr-4G platinum-iridium electrode coating is a proprietary innovation that has not been used in any previously cleared CGM device in the QKQ product code space, based on our review of publicly available 510(k) summaries. While the underlying electrochemical sensing principle (enzymatic glucose oxidase reaction on a metallic electrode) is the same, the introduction of iridium into the electrode alloy represents a material change that could, in FDA's view, raise questions about biocompatibility (particularly with respect to extractable iridium species), electrochemical performance (sensor drift characteristics, interference susceptibility), and long-term signal stability over the 14-day wear period. We expect FDA to acknowledge that this difference does not fundamentally alter the technological basis of the device but to request additional supporting data to demonstrate that the PtIr-4G electrode does not introduce new safety or effectiveness risks relative to the platinum electrode in K211034.
Sensor Wear Duration (14 days vs. 10 days). This difference is effectively bridged by the secondary predicate K230489 (NX-700), which has an identical 14-day wear duration. We do not anticipate significant FDA pushback on this point, particularly given that the NX-700 clearance established FDA's comfort with 14-day enzymatic electrochemical CGM sensors within the QKQ product code. FDA may, however, request sensor performance data specifically addressing the final days of the 14-day wear period (e.g., Days 11–14) to confirm that accuracy and signal quality are maintained throughout the full wear duration.
Reader Form Factor (Arm-Worn vs. Handheld). The Lumina R7's arm-worn integrated reader/transmitter represents a design change relative to the NX-500's handheld receiver, but this is not a change to the fundamental scientific technology of the device. The R7 performs the same functions as the NX-500 reader — receiving BLE-transmitted glucose data, displaying real-time glucose values and trends, and generating configurable alerts — but in a different physical form factor. FDA typically addresses form-factor changes through human factors and usability testing rather than treating them as predicate-breaking differences. We expect FDA to confirm that this difference is appropriately addressed through a human factors validation study (consistent with FDA's February 2016 guidance on human factors and the IEC 62366-1 standard) rather than through additional clinical or bench performance testing.
BLE Version (5.2 vs. 5.0). The upgrade from BLE 5.0 to BLE 5.2 is an incremental wireless communication protocol change that does not alter the fundamental technology of the device or the nature of the data transmitted between the sensor and reader. We do not anticipate that FDA will treat this difference as a substantial equivalence concern. However, as discussed in greater detail in Section III.F below, FDA may use the BLE 5.2 implementation as a basis for heightened cybersecurity documentation expectations, given that BLE 5.2 introduces new features (such as LE Audio and isochronous channels) that may present different cybersecurity considerations.
Proprietary Algorithm (LuminaPredict v2.1 vs. NovaSense Algorithm). Different proprietary algorithms for glucose calibration, signal processing, and predictive alerting are a standard feature of CGM devices and are routinely addressed in 510(k) substantial equivalence analyses through performance data demonstrating equivalent or superior accuracy (typically measured by MARD and error grid analysis). FDA does not generally require algorithm-level equivalence between a subject device and its predicate; rather, the Agency evaluates whether the subject device's algorithm produces clinically equivalent outputs. We expect FDA to accept this approach, provided that the LUMINA-US clinical study generates adequate performance data. The primary area of potential concern is whether the LuminaPredict v2.1 algorithm incorporates any machine learning ("ML") or artificial intelligence ("AI") components, which would trigger additional documentation requirements under FDA's AI/ML-based Software as a Medical Device ("SaMD") framework. This point requires confirmation from Dr. Aisling Burke, as discussed in Section VII below.
III.A.2. Risk Assessment
Risk of predicate rejection: LOW. Both K211034 and K230489 are well-established CGM predicates within the QKQ product code space. Both were cleared through the 510(k) pathway, both are Class II devices under the same regulatory classification, and both share the same fundamental technological characteristics as the Lumina System. The split-predicate approach is a recognized and accepted CDRH strategy, and the combination of K211034 (establishing the fundamental technology platform) and K230489 (bridging the wear duration) provides a robust substantial equivalence argument. We have identified no publicly available precedent in which CDRH rejected a QKQ predicate under analogous circumstances.
Risk that FDA treats the PtIr-4G electrode coating as raising a "new question of safety or effectiveness": MODERATE. The PtIr-4G technology is a proprietary advancement with no direct precedent in the QKQ product code space. While the underlying electrochemical sensing principle is unchanged, FDA's reviewing team may view the introduction of a new electrode alloy material as a change that warrants additional supporting data beyond what would normally be required for a straightforward 510(k) submission. Specifically, FDA may request: (a) comparative bench data demonstrating equivalent or superior sensor drift performance of PtIr-4G versus conventional platinum over a 14-day wear period; (b) interference testing (e.g., ascorbic acid, acetaminophen, hydroxyurea) specific to the PtIr-4G electrode; (c) targeted biocompatibility studies addressing the iridium component (particularly extractables/leachables analysis under ISO 10993-18); and (d) any available accelerated stability data for the PtIr-4G electrode. These are data requests that can be addressed through additional bench testing and targeted biocompatibility studies, not through a fundamentally different regulatory pathway. The risk is therefore characterized as moderate in terms of likelihood but manageable in terms of impact.
III.A.3. Contingency Strategy — If FDA Rejects Primary Predicate
In the event that FDA determines K211034 (NovaSense GlucoTrack NX-500) is not an appropriate primary predicate — for example, due to a determination that the combined weight of the five identified technological differences, considered together rather than individually, raises different questions of safety or effectiveness — the first contingency is to redesignate K230489 (NovaSense GlucoTrack NX-700) as the primary predicate device. The NX-700 shares the Lumina System's 14-day wear duration, and the remaining differences (electrode material, BLE version, reader form factor, algorithm) would need to be individually addressed through performance data. This approach is viable and has the advantage of reducing the number of predicate differences to four (eliminating the wear duration difference). However, it narrows the flexibility of the substantial equivalence argument by placing all of the SE burden on a single predicate, and it may invite additional scrutiny of the remaining differences.
If FDA were to reject both NovaSense predicates — a scenario we assess as unlikely but not impossible, particularly if FDA views the PtIr-4G electrode chemistry as fundamentally distinct from all previously cleared CGM electrode technologies — Solstice would need to identify alternative predicate devices within the QKQ product code. Amy Whitfield and Sean Morrissey have conducted a comprehensive review of the FDA 510(k) database for all QKQ clearances issued in the past five years (2020–2025). This review identified three additional QKQ clearances from manufacturers other than NovaSense; however, all three devices employ conventional platinum or platinum-based electrode technologies. None of these alternative devices would provide a closer technological match to the Lumina System's PtIr-4G electrode than the NovaSense predicates. Accordingly, if both NovaSense predicates are rejected, the identification of a suitable alternative predicate becomes significantly more challenging.
In the extreme scenario where no suitable predicate device can be identified, Solstice would face two regulatory pathway options: (a) a De Novo classification request under Section 513(f)(2) of the FD&C Act, which would establish a new classification for the Lumina System and create a new regulatory pathway for future similar devices; or (b) a PMA submission under Section 515 of the FD&C Act, which would require a significantly more extensive clinical evidence package and impose a substantially higher regulatory burden. The cost, timeline, and strategic implications of each of these alternative pathways are discussed in Section III.B below.
We assess the probability of needing to pursue either of these alternative pathways as low. Nevertheless, the contingency analysis is provided to ensure that Solstice and the engagement team are prepared for all reasonably foreseeable outcomes of the Pre-Sub meeting.
III.B. Contingency: PMA vs. 510(k) Pathway
III.B.1. Likelihood of FDA Requiring PMA
We assess the probability that FDA will require Solstice to pursue a PMA submission for the Lumina CGM System as low. This assessment is based on the following considerations:
First, the Lumina CGM System falls within a well-established Class II product code (QKQ) with multiple cleared predicate devices. The 510(k) pathway has been the standard regulatory route for CGM devices in this product code category, and FDA has not required PMA for any enzymatic electrochemical CGM system cleared under QKQ in recent years.
Second, FDA's March 2024 reclassification of glucose test systems under 21 C.F.R. § 862.1355 reaffirmed the Class II classification with special controls for CGM devices. This reclassification was a deliberate regulatory action by CDRH to modernize and consolidate the regulatory framework for glucose monitoring devices, and it reflects FDA's institutional view that the safety and effectiveness of these devices can be adequately assured through 510(k) clearance with special controls rather than through PMA.
Third, a PMA requirement would be legally supportable only if FDA determined that the Lumina System raises entirely new questions of safety or effectiveness that cannot be resolved through the 510(k) substantial equivalence framework. While the PtIr-4G electrode material is a novel element, it represents an incremental material innovation within an established sensing technology platform — not a fundamentally new device concept or therapeutic modality.
The most realistic trigger for a PMA demand would be a determination by FDA that the LuminaPredict v2.1 algorithm introduces a novel therapeutic function that goes beyond the diagnostic and monitoring functions of existing cleared CGM devices. For example, if the algorithm's outputs were designed to directly control an insulin pump or automated insulin delivery system without user intervention (i.e., a "closed-loop" insulin delivery function), FDA would likely view the device as having a different intended use than existing CGM predicates, potentially requiring PMA or a new De Novo classification. As currently designed, the Lumina System provides glucose data display, trend information, and configurable hypo/hyperglycemia alerts only; it does not incorporate any automated insulin dosing functionality and is not designed to directly interface with insulin delivery systems. This distinction must be clearly articulated in the Pre-Sub meeting package and reinforced during the Pre-Sub meeting itself.
III.B.2. Contingency if PMA is Required
Should FDA nonetheless determine that a PMA is required, the implications for Solstice would be substantial:
Cost. A PMA pathway would entail estimated additional costs of $2,000,000 to $5,000,000 in additional clinical data generation (including an expanded clinical study with a larger sample size, longer follow-up duration, and potentially a randomized controlled design), regulatory consulting fees, and internal resource allocation. The PMA application user fee for FY2025 is approximately $425,000 for a standard PMA application (based on the most recent FDA user fee schedule published under MDUFA V), compared to approximately $22,000 for a standard 510(k) submission.
Timeline. The PMA pathway would extend the regulatory timeline by an estimated 12 to 18 months beyond the current Q3 2026 clearance target, potentially pushing FDA approval to Q1 2028 or later. PMA review times have averaged approximately 180 FDA review days in recent years, but the total elapsed time — including FDA requests for additional information, advisory committee scheduling (if applicable), and manufacturing inspection — is typically 18 to 24 months from PMA submission to approval.
Commercial Impact. A PMA requirement would require Solstice to fundamentally reevaluate the commercial viability of its U.S. market entry. Dr. Eoin Gallagher (CEO, Solstice Biomedical Devices Ltd.) and the board of directors would need to make a go/no-go decision based on the revised cost-benefit analysis, considering the additional investment required, the delayed revenue timeline, and the competitive landscape in the U.S. CGM market.
De Novo as Intermediate Option. An intermediate regulatory pathway — the De Novo classification request under Section 513(f)(2) — could be explored if FDA determines that the Lumina System is a novel device for which there is no legally marketed predicate but which presents a low-to-moderate risk profile that can be adequately addressed through general and special controls. The De Novo pathway has several advantages over PMA: it does not require the extensive clinical evidence package of a PMA, the review timeline is typically 150 FDA review days (compared to 180 for PMA), and the FY2025 De Novo user fee is approximately $134,000 (significantly less than the PMA fee). However, the De Novo pathway also has limitations: it requires FDA to establish a new device classification and identify appropriate special controls, which can be a complex and time-consuming process, and it provides less regulatory predictability than either the 510(k) or PMA pathways.
Recommendation. Our strong recommendation is to emphasize the robust substantial equivalence argument in the Pre-Sub meeting package, clearly articulate the split-predicate framework, and proactively address the five identified technological differences with supporting data and analysis. The goal should be to foreclose any PMA discussion at the Pre-Sub meeting by demonstrating that the Lumina System fits squarely within the established QKQ 510(k) framework.
III.C. Question 2: Clinical Study Design and Endpoints
III.C.1. Most Likely FDA Position
Based on our review of the September 2023 iCGM guidance, recent 510(k) clearances for CGM devices, and established CDRH expectations for clinical studies supporting CGM 510(k) submissions, we anticipate that FDA will find the overall LUMINA-US clinical study design to be generally appropriate for a 510(k) CGM submission. Specifically:
Study Design. The proposed prospective, single-arm, multicenter study design with 350 subjects (175 with Type 1 diabetes and 175 with Type 2 diabetes) enrolled at 6 U.S. clinical sites is consistent with the study designs accepted by CDRH in recent CGM 510(k) clearances. FDA's September 2023 iCGM guidance does not mandate a specific study design but identifies key performance criteria and study parameters; the LUMINA-US design is aligned with these parameters. The single-arm design — in which the Lumina System's glucose readings are compared against a laboratory reference standard rather than against a comparator device — is the standard approach for CGM accuracy studies and is unlikely to be challenged.
Primary Endpoint — MARD. The proposed primary endpoint of Mean Absolute Relative Difference ("MARD") ≤ 9.0% is aligned with current FDA expectations for iCGM-level accuracy performance. The 9.0% MARD threshold is consistent with the special controls established in the iCGM De Novo classification order (DEN170088) and has been adopted by FDA as a key performance benchmark for integrated CGM systems. We do not anticipate that FDA will object to this primary endpoint or request a different MARD threshold.
Secondary Endpoint — Consensus Error Grid. The proposed secondary endpoint requiring ≥ 99% of paired glucose readings to fall within Zones A and B of the Clarke Error Grid is consistent with CDRH precedent and established expectations for CGM accuracy performance. We note, however, that FDA may also request a Surveillance Error Grid ("SEG") analysis as an additional secondary endpoint. The SEG is a more recently developed error grid framework that provides a more granular assessment of clinical risk associated with glucose measurement errors, and FDA has begun referencing the SEG in its review of CGM submissions. While a SEG analysis is not yet uniformly required, Solstice should be prepared to include it as a secondary endpoint.
Sub-Analyses. We anticipate that FDA will request, either in its Pre-Sub feedback or during the Pre-Sub meeting discussion, specific sub-analyses of the MARD data, including:
(a) MARD broken out by glucose range — specifically, performance in the hypoglycemic range (< 70 mg/dL), the euglycemic range (70–180 mg/dL), and the hyperglycemic range (> 180 mg/dL). This sub-analysis is particularly important for iCGM classification, as the iCGM special controls include specific accuracy thresholds for hypoglycemic and hyperglycemic readings.
(b) MARD by day of sensor wear — for example, Days 1–3, Days 4–7, Days 8–10, and Days 11–14. This sub-analysis is relevant to the 14-day wear duration claim and will allow FDA to assess whether sensor accuracy and signal quality are maintained throughout the full 14-day wear period, including the final days when sensor drift and biofouling are most likely to affect performance. This sub-analysis is particularly important given the novel PtIr-4G electrode material, which Solstice claims offers improved resistance to sensor drift.
(c) Performance during challenging physiological conditions, including rapid glucose rate of change (> 2 mg/dL/min), post-meal glucose excursions, and during exercise. These conditions are known to affect CGM accuracy and are areas where FDA expects demonstration of adequate performance.
User Labeling Comprehension Study. The inclusion of a user labeling comprehension study as part of the clinical protocol is appropriate and expected. FDA routinely requires evidence that users can correctly understand and follow the device labeling instructions, and this is particularly important for the Lumina System given the arm-worn R7 reader form factor (which differs from the handheld readers that users may be familiar with from other CGM systems) and the self-insertion procedure for the C14 sensor.
III.C.2. Risk of Additional Endpoint Demands
Pediatric Labeling Requirement: MODERATE RISK. There is a moderate risk that FDA will request a specific pediatric labeling plan or require Solstice to provide a detailed justification for the 18+ age restriction on the intended use population. The CGM market has seen a significant trend toward adolescent and pediatric use in recent years, with several CGM systems now cleared for use in patients as young as 2 years of age. FDA may question whether limiting the initial indication to adults is clinically justified or whether it reflects a commercial decision by Solstice to defer pediatric labeling. We recommend that Solstice prepare a written rationale for the initial adult-only indication, emphasizing that: (i) the LUMINA-US clinical study is designed to generate data in the adult population that is directly relevant to the 510(k) submission; (ii) pediatric CGM use involves different clinical considerations (e.g., smaller body habitus, different glucose variability patterns, caregiver-dependent use) that warrant separate evaluation; and (iii) Solstice intends to pursue a pediatric indication expansion through a post-market clinical study and a supplemental 510(k) submission following initial clearance.
Comparator Arm or Head-to-Head Data: LOW-TO-MODERATE RISK. There is a low-to-moderate risk that FDA will request a comparator arm or head-to-head data comparing the Lumina System's performance against a predicate CGM device. This type of request is atypical for 510(k) CGM submissions — the standard approach is to evaluate accuracy against a laboratory reference standard (typically the Penrose Software or equivalent), not against a competing commercial device. However, in isolated cases, FDA has requested comparator data when the technological differences between the subject device and the predicate are significant enough to raise questions about relative performance. Should this request arise, Solstice should be prepared to argue that accuracy measured against a laboratory reference standard is the appropriate and established methodology per the September 2023 iCGM guidance, and that head-to-head comparison against a predicate device does not provide meaningful additional safety or effectiveness information.
Clinical Study Population Diversity: LOW-TO-MODERATE RISK. FDA may ask about the demographic composition of the LUMINA-US study population, with particular attention to adequate representation of diverse skin types (Fitzpatrick skin types IV–VI) and diverse body habitus (including individuals with higher BMI). This request would be driven by FDA's recent emphasis on the importance of ensuring that medical devices perform adequately across the full range of the intended use population, particularly for devices with skin-contacting components where adhesion, skin irritation, and optical/sensor performance may vary across skin types. For the Lumina System, the primary concern is the R7 reader's adhesive contact with the skin and the C14 sensor's subcutaneous insertion characteristics across different body compositions. Solstice should ensure that the LUMINA-US study enrollment plan includes adequate representation of diverse demographics and should be prepared to discuss this aspect of the study design at the Pre-Sub meeting.
III.C.3. Contingency — If FDA Demands Additional Endpoints
If FDA requires additional endpoints beyond those currently specified in the LUMINA-US study protocol, Solstice should evaluate whether these additional endpoints can be incorporated into the existing Protocol SBD-LUM-CL-001 without increasing enrollment beyond the planned 350 subjects. Many of the sub-analyses that FDA is likely to request (glucose range-specific MARD, day-of-wear MARD, challenging conditions performance) can be performed on the same dataset collected from the same 350 subjects without requiring additional study procedures or additional subjects. The protocol should be designed with sufficient flexibility to accommodate these analyses, and the statistical analysis plan should pre-specify the methodology for each sub-analysis.
If FDA requires a Surveillance Error Grid analysis, this can be performed on the same paired glucose data (device reading vs. reference standard) that is collected for the primary MARD endpoint and the Clarke Error Grid analysis. No additional data collection is needed; only additional statistical analysis. This should be straightforward to accommodate.
If, however, FDA requires a pediatric sub-study as a condition for initial clearance — rather than accepting a post-market commitment — this would represent a significant scope expansion with substantial cost and timeline implications. A pediatric sub-study would require: (i) separate pediatric-specific clinical protocol development and IRB review; (ii) identification and qualification of pediatric clinical sites with appropriate expertise; (iii) enrollment of a pediatric cohort (estimated 50–100 additional subjects, depending on FDA's requirements); and (iv) additional time for enrollment, data collection, and analysis. We estimate that a pediatric sub-study requirement would extend the overall regulatory timeline by approximately 6 to 9 months and add an estimated $1,500,000 to $2,000,000 in additional clinical study costs. Our recommendation is to resist a pediatric requirement for the initial clearance and to propose a post-market pediatric study with a defined timeline and milestones.
We note that the current clinical study budget of $4,200,000 (based on 350 subjects at an average cost of $12,000 per subject) through Veltrano Clinical Research Organization Inc. has a limited contingency margin. Any significant protocol amendment that increases the number of study procedures per subject, extends the per-subject follow-up period, or adds analytical complexity would likely require a budget revision with Veltrano. This budget constraint should be factored into any negotiations with FDA regarding additional clinical requirements.
III.D. Question 3: IDE Exemption Applicability (NSR Determination)
III.D.1. Most Likely FDA Position
We anticipate that FDA will likely concur that the LUMINA-US clinical study qualifies for the Investigational Device Exemption ("IDE") exemption under 21 C.F.R. § 812.2(c)(3), which exempts diagnostic devices from the full IDE requirements, provided that the study meets the criteria for a non-significant risk ("NSR") investigation. Our assessment is based on the following analysis:
The Lumina C14 sensor is a subcutaneously implanted device, which places it in a category that requires careful NSR/SR evaluation. However, the C14 is not a life-sustaining or life-supporting implant in the traditional regulatory sense — it is a diagnostic monitoring sensor that measures interstitial glucose concentration but does not deliver any therapeutic substance or perform any therapeutic function. The LUMINA-US study involves diagnostic monitoring only; subjects will continue their existing diabetes management regimens and will not use the Lumina System for insulin dosing decisions during the study (instead relying on their existing glucose monitoring systems for treatment decisions). This study design is consistent with the NSR determination criteria set forth in 21 C.F.R. § 812.3(m), which defines a significant risk device as one that, among other things, is an implant presenting a potential for serious risk to health, safety, or welfare. While the C14 is subcutaneously implanted, the risk profile is substantially different from traditional significant risk implants (e.g., cardiac pacemakers, orthopedic implants) given the short implantation duration (14 days), the minimally invasive insertion procedure, and the diagnostic-only function.
FDA may note during the Pre-Sub meeting that the NSR determination ultimately rests with the reviewing Institutional Review Board ("IRB") — in this case, Clearwater Institutional Review Board LLC (IRB Registration No. IRB-2019-04821) — and that FDA does not formally "approve" or "endorse" NSR determinations in the Pre-Sub context. FDA's role at the Pre-Sub stage is limited to indicating whether it would likely agree with an IRB's NSR conclusion, based on the study design and the device's risk profile. We expect FDA to provide a generally supportive indication, while noting this procedural distinction.
There is a nuance that warrants attention: the C14 sensor is subcutaneously implanted for 14 days, and some IRBs have historically classified subcutaneous implantable sensors — even short-duration diagnostic sensors — as significant risk devices requiring a full IDE. This classification tendency varies among IRBs, and Clearwater IRB's own track record with CGM studies is not publicly available. If Clearwater IRB determines that the LUMINA-US study involves a significant risk device, Solstice would be required to submit a full IDE application to FDA before commencing the study, regardless of FDA's Pre-Sub indication.
III.D.2. Contingency — If IDE Is Required
If FDA indicates during the Pre-Sub meeting that the LUMINA-US study is likely to be classified as significant risk, or if Clearwater IRB independently makes an SR determination, Solstice should be prepared to submit an IDE application (Form FDA 3514) to FDA. The key implications of this contingency are as follows:
Review Timeline. FDA review of an IDE application is 30 calendar days from receipt of a complete submission. This is a statutory timeline and is generally adhered to by CDRH for diagnostic device IDE applications.
Application Requirements. The IDE application would require submission of the complete clinical study protocol (SBD-LUM-CL-001), all investigator agreements and curricula vitae, IRB documentation (including the SR determination and any IRB conditions), the clinical monitoring plan, and the device description and manufacturing information. Much of this documentation is already in preparation for the Pre-Sub package and the study initiation process, so the incremental regulatory burden of compiling an IDE application is manageable, though not trivial.
Timeline Impact. If an IDE is required, the study commencement date would shift from the currently targeted September 15, 2025 to approximately October 15, 2025 (accounting for IDE preparation, submission, and the 30-day FDA review period). This one-month delay would cascade through the downstream timeline, potentially impacting the study completion date (shifting from March 15, 2026 to approximately April 15, 2026), the 510(k) submission date (shifting from April 30, 2026 to approximately May 31, 2026), and the target clearance date (shifting from Q3 2026 to late Q3 or early Q4 2026). While this delay is not catastrophic, it narrows the margin of error in the overall timeline and reduces the buffer available for addressing any unforeseen complications during the clinical study.
Recommendation. We recommend that Solstice prepare a "shadow" IDE application in parallel with the primary NSR pathway. This shadow application should be drafted to a substantially complete state by August 31, 2025, so that it can be finalized and submitted on short notice if Clearwater IRB makes an SR determination. The cost of preparing the shadow IDE application is modest (estimated at $15,000–$25,000 in additional regulatory consulting fees from Pinnacle), and the optionality it provides is well worth the investment.
III.E. Question 4: Software Documentation Requirements and Level of Concern
III.E.1. Most Likely FDA Position
The software documentation question is an area where FDA's expectations have evolved significantly in recent years, and Solstice must ensure that its approach reflects the current regulatory framework rather than historical practices. Our assessment of FDA's likely position is as follows:
Software Safety Classification. FDA will likely agree that the software components of the Lumina System warrant a classification of "Major" under the legacy "level of concern" framework, given the clinical significance of the glucose readings and alerts provided by the system. However, we note that FDA's current framework — as reflected in the September 2023 guidance documents and CDRH's evolving review practices — has shifted toward using the IEC 62304:2006/A1:2015 software safety classification system rather than the legacy FDA "level of concern" categorization. Under IEC 62304, the software components would likely be classified as either Class B (software that could contribute to a hazardous situation resulting in non-serious injury) or Class C (software that could contribute to a hazardous situation resulting in death or serious injury), depending on FDA's assessment of the hazard potential associated with inaccurate glucose readings or missed alerts. The distinction between Class B and Class C has significant implications for the required software development documentation: Class C requires the most comprehensive documentation, including detailed design and detailed design verification, while Class B has somewhat reduced requirements. We recommend that Solstice prepare documentation at the Class C level to ensure compliance regardless of FDA's classification determination.
Documentation Expectations. FDA will expect a comprehensive software documentation package for all four software components of the Lumina System. This documentation should include, at a minimum:
• Software requirements specification (SRS) for each component
• Software architecture design document
• Detailed software design document
• Software development plan and software development environment documentation
• Software verification and validation (V&V) protocols and test reports
• Requirements traceability matrix linking requirements to design elements, verification tests, and risk controls
• Unresolved software anomaly analysis (open bug list with risk assessment and disposition)
• Software revision history and change control documentation
• Software of Unknown Provenance (SOUP) analysis for any third-party software components or libraries
This documentation must be provided for each of the four software components: Lumina Firmware v3.4.2 (C14 sensor), Lumina Reader OS v2.1.0 (R7 reader), LuminaPredict Algorithm v2.1, and Lumina Mobile App v1.0.3.
LuminaPredict Algorithm Scrutiny. The LuminaPredict v2.1 algorithm will receive heightened scrutiny from FDA's software review team, given its central role in converting raw electrochemical signals into clinically actionable glucose values and trend predictions. FDA will likely request detailed documentation of the algorithm's design rationale, training data (if applicable), validation datasets, and performance metrics. Critically, if LuminaPredict v2.1 incorporates any machine learning ("ML") or artificial intelligence ("AI") components — even as minor sub-functions within the overall algorithm — this would trigger significant additional documentation requirements under FDA's AI/ML-based SaMD framework, including documentation of the training and validation data provenance, the algorithm architecture and hyperparameters, performance across relevant subpopulations, and a predetermined change control plan for any planned algorithm modifications. Dr. Aisling Burke (CTO, Solstice Biomedical Devices Ltd.) must confirm by May 15, 2025 whether any AI/ML components are incorporated into LuminaPredict v2.1, so that the Pre-Sub package can be prepared accordingly.
Software Bill of Materials (SBOM). An SBOM is mandatory for the 510(k) submission under the September 2023 cybersecurity guidance. The SBOM must enumerate all commercial, open-source, and off-the-shelf software components used in the Lumina System, including component name, version, and manufacturer/source. The SBOM should be prepared in a machine-readable format (e.g., SPDX or CycloneDX) as well as a human-readable format for inclusion in the 510(k) submission.
Mobile Application Regulatory Status. FDA may raise a question regarding the regulatory status of the Lumina Mobile App v1.0.3 — specifically, whether the mobile application is itself a regulated medical device or whether it falls within the enforcement discretion framework for mobile medical applications established under Section 520(o) of the FD&C Act. Given that the Lumina Mobile App displays glucose values, generates alerts, and executes portions of the LuminaPredict algorithm, it is likely that FDA will consider the app to be a regulated device (specifically, a device accessory) rather than an app within the scope of enforcement discretion. Solstice should be prepared to include the mobile application within the scope of the 510(k) submission and provide the full suite of software documentation for the app.
III.E.2. Identified Gaps
Sean Morrissey prepared a Software Documentation Checklist and Gap Analysis (dated April 28, 2025) based on a preliminary review of the software documentation packages provided by Solstice's engineering team. That analysis identified the following potential gaps that must be addressed before the 510(k) submission:
(a) The software architecture documentation for the Lumina Mobile App v1.0.3 is incomplete — specifically, the architecture design document does not adequately describe the interface between the mobile app and the cloud synchronization backend, and the data flow diagrams do not clearly illustrate the handling of glucose data at rest on the mobile device versus data in transit to the cloud server.
(b) The requirements traceability matrix for the Lumina Reader OS v2.1.0 has incomplete linkages between several software requirements and their corresponding verification test cases. Approximately 12 software requirements (out of approximately 340 total requirements) lack documented verification test case references. While these gaps may reflect documentation omissions rather than actual testing gaps, they must be resolved to present a complete and defensible traceability matrix to FDA.
These gaps were communicated to Dr. Aisling Burke and Fiona Drennan (Quality Assurance Director) on May 1, 2025, with a request for remediation plans by May 20, 2025. We recommend that these gaps be fully resolved before the Pre-Sub meeting package is finalized to ensure that Solstice can credibly represent to FDA that its software documentation is substantially complete.
III.F. Question 5: Cybersecurity Premarket Submission Expectations
III.F.1. Most Likely FDA Position
Cybersecurity is an area of intensifying FDA focus, and the Lumina System's BLE-connected architecture places it squarely within the scope of CDRH's cybersecurity expectations. We anticipate that FDA will expect full compliance with the September 2023 final guidance "Cybersecurity in Medical Devices: Quality System Considerations and Content of Premarket Submissions," which superseded the 2014 premarket cybersecurity guidance and established significantly more detailed and prescriptive requirements.
The key elements that FDA will expect in the 510(k) submission's cybersecurity documentation package include:
• Threat Model: A comprehensive threat model using an established methodology such as the STRIDE (Spoofing, Tampering, Repudiation, Information Disclosure, Denial of Service, Elevation of Privilege) framework. The threat model should address all communication interfaces (BLE 5.2 between C14 and R7, BLE between R7 and mobile device, internet communication between mobile app and cloud backend) and all potential threat actors.
• Software Bill of Materials (SBOM): As discussed in Section III.E.1 above, a complete SBOM enumerating all software components, versions, and sources.
• Cybersecurity Risk Assessment: A documented cybersecurity risk assessment that identifies, evaluates, and mitigates cybersecurity risks across the device lifecycle, including risks associated with unauthorized access, data integrity compromise, device malfunction due to cyber attack, and unauthorized modification of device software.
• Vulnerability Testing Results: Documentation of security testing, including static code analysis, dynamic analysis, fuzz testing, and penetration testing results for all software components and communication interfaces.
• Software Update and Patch Management Plan: A documented plan for delivering software updates and security patches throughout the device's lifecycle, including the mechanism for update delivery, the validation process for updates, and the communication plan for notifying users and healthcare facilities of available updates.
• Secure Authentication Protocols: Documentation of the authentication mechanisms used for BLE 5.2 communication between the C14 sensor and the R7 reader, and between the R7 reader and the mobile application, including the cryptographic protocols employed.
• Data Encryption: Documentation of data encryption methods for data in transit (between device components and between the mobile app and cloud backend) and data at rest (stored glucose data on the R7 reader and mobile device).
• End-of-Life / End-of-Support Plan: A documented plan addressing the end of cybersecurity support for the device, including the process for notifying users when security updates will no longer be provided.
FDA may also direct particular attention to the BLE 5.2 implementation, as BLE 5.2 is the most current iteration of the BLE standard and introduces several new features (including LE Audio, Enhanced Attribute Protocol, and isochronous channels) that may present cybersecurity considerations not addressed in earlier BLE versions. Specifically, FDA may ask about the specific BLE 5.2 security features leveraged by the Lumina System — such as LE Secure Connections mode and AES-CCM (Advanced Encryption Standard, Counter with CBC-MAC) encryption — and whether Solstice has assessed and mitigated known BLE vulnerabilities that have been publicly disclosed in recent years, including the KNOB (Key Negotiation of Bluetooth) attack, the BLESA (Bluetooth Low Energy Spoofing Attack) attack, and the BLURtooth vulnerability.
Additionally, FDA may request evidence that Solstice has established a coordinated vulnerability disclosure ("CVD") policy — a mechanism by which external cybersecurity researchers can report vulnerabilities to Solstice in a structured and responsible manner — and evidence that Solstice has engaged with the cybersecurity research community regarding the Lumina System's security architecture.
III.F.2. Risk Assessment
Risk of cybersecurity-related deficiency finding: MODERATE. Cybersecurity has been one of the most frequent areas of FDA 510(k) "Refuse to Accept" ("RTA") decisions in 2024 and 2025. CDRH has publicly stated that incomplete cybersecurity documentation is a leading cause of RTA decisions for connected medical device submissions, and multiple CGM and connected device 510(k) submissions have been issued RTA letters specifically for insufficient SBOM documentation, inadequate or missing threat modeling, or failure to address the September 2023 guidance requirements. An RTA decision does not constitute a substantive review finding — it simply means that FDA determined the submission was incomplete and woul