Clinical Evidence to Support US Food and Drug Administration Review of New Medical Technology in Pulmonary, Sleep, and Critical Care Medicine Between 2014 and 2024

Abstract

Background

Chest medicine relies extensively on medical devices that comprise medical technology, artificial intelligence, software applications, and digital health; however, a comprehensive assessment of the evidence supporting US Food and Drug Administration (FDA) approval/clearance of these devices is lacking.

Research Question

What are the trends and evidence related to new medical devices across chest medicine?

Study Design and Methods

We performed a scoping review according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews guidelines to assess medical devices approved or cleared by the FDA between June 1, 2014, and June 30, 2024.

Results

A total of 790 new devices were FDA-approved or cleared in the last decade. Of these, 98% of devices were cleared through a 510k pathway, indicating they were deemed similar to other products already on the market. Almost all Premarket Approval devices and one-half of De Novo devices were approved/classified using clinical data that measured objective clinical primary outcomes rather than patient-reported primary outcomes.

Interpretation

Our results show that approval for new medical devices in pulmonary, sleep, and critical care medicine largely depends on establishing equivalence to existing devices. Clinical trials are generally reserved for higher-risk devices that are interventional in nature or monitor the patient’s status.

Take-Home Points.

Research Question: What are the trends and evidence related to new medical devices across chest medicine?

Results: A total of 790 new devices were FDA-approved or cleared in the last decade for pulmonary, sleep, and critical care medicine.

Interpretation: Our results show that most new medical devices do not require new clinical trials to reach the market, and clinical trials are generally reserved for higher-risk interventional devices.

New medical technology creates avenues for delivering high-quality chest medicine that broadly comprises pulmonary, sleep, and critical care medicine. From artificial intelligence software to promote healthy sleep to physical devices to facilitate bronchoscopic lung volume reduction, novel technologies provide a myriad of benefits to patients across the spectrum of chest medicine.^1,2 The US Food and Drug Administration (FDA) regulates medical devices, which comprise traditional medical technology, artificial intelligence, software, and digital health platforms. Since the COVID-19 pandemic, the push for remote monitoring technologies and virtual care has created new opportunities for innovation even in traditional medical device paradigms—solidifying the key role that new devices will play in the future.³

Physicians report being unfamiliar with how medical devices are regulated.^4,5 Physician perceptions on device risk, justified or not, impact their use in practice.⁶ Addressing these needs is therefore important because successful innovation hinges on partnerships among industry, medical professional societies, and clinical experts to anticipate gaps between the clinical data package needed to support regulatory approval of new devices and the full set of clinical data that are important for their successful adoption in practice. The regulatory view of device regulation anchors on risk assessment, and does not intend under government purview to encapsulate the totality of clinical data needed to support successful clinical adoption in full. To reach the market, most devices are determined to be low risk and can be FDA-cleared through the 510(k) process as substantially equivalent to an established product.⁷ For example, Auris Monarch (Johnson & Johnson MedTech) and Intuitive Ion (Intuitive Surgical) navigation systems did not require clinical data to support their FDA clearance. Recognizing that pulmonologists perform bronchoscopy for various reasons beyond the scope of a single randomized controlled trial end point, regulatory frameworks are largely designed to defer pragmatic clinical evidence generation to medical professionals who will use these platforms.⁸ In contrast to low-risk devices, devices that pose moderate or high risk to patients require clinical data to demonstrate probable benefits outweigh probable risks and are evaluated through processes such as the Premarket Approval (PMA) pathway.⁹ Some low-risk devices have no established predicate device on which to establish FDA clearance, and this gap is resolved by availability of the De Novo pathway that often requires clinical data to support labeling claims for new indications or disease states.

We performed a scoping review evaluating the availability and rigor of clinical evidence to support medical device approvals and clearances across chest medicine over the last decade. Gaps in clinical evidence required by regulators inherently may reflect key opportunities for pragmatic trials to support clinical adoption in a successful innovation ecosystem.^10,11

Study Design and Methods

We reviewed technologies directly from FDA medical device databases for 510(k) Premarket Notification, De Novo, and PMA.¹² Searches were limited to the Anesthesiology Advisory Committee, which oversees device applications across most pulmonary care fields applicable to chest medicine. Each database was searched independently for devices that were FDA-cleared or approved between June 1, 2014, and June 30, 2024. This study was conducted in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews guidelines,¹³ leveraging descriptive (noncomparative) analysis, absent any formal hypothesis testing. To further characterize the quality of evidence supporting new technologies, included trials were reviewed in a targeted manner, noting Consolidated Standards of Reporting Trials (CONSORT) adherence, primary end points, blinding, and statistical plan.

Review

Two research team members (J. T. G. and J. A. B.) independently and manually reviewed all device applications. Devices receiving Humanitarian Device Exemption and those indicated for use outside pulmonary, sleep, and critical care medicine were excluded. We included original applications for new devices between 2014 and 2024, given our focus on understanding current FDA thinking regarding adequacy of clinical trial data to support new device applications. We excluded 510(k) and PMA modifications on labeling, manufacturing process, device design, and software updates. Such modifications are clearly linked to an original parent application in FDA-approval letters.

Data Extraction and Analysis of Regulatory Review

Extracted information for devices included name, manufacturer, registration number, and decision date. Domains of interest for this scoping review included the following: (1) indication for the device, (2) presence and quality of clinical trial data, and (3) adverse events. Results were stratified on the 510(k), De Novo, and PMA regulatory pathways available to industry sponsors based on the level of risk that investigational devices pose to patients (class I: low risk; class II: low-to-moderate risk; and class III: high risk or implantable devices). Device indications were classified as: general (eg, software for anesthesia machine), delivery (eg, high-flow oxygen delivery device), monitoring (eg, carbon monoxide gas analyzer), or intervention (eg, implanted phrenic nerve stimulator) while noting the presence of FDA-reviewed clinical trial data in the approval/summary letter. For applications containing clinical data, we extracted data from published clinical trials and clinicaltrials.gov to assess the quality of evidence in the study according to the CONSORT statement.¹⁴ Data not reported in FDA summaries, published clinical trials, or clinicaltrials.gov were recorded as not reported and excluded from denominators for relevant percentages. We focused our analysis on primary outcomes that are often determined jointly between industry sponsors of investigational products and FDA review teams, rather than secondary and exploratory outcomes that are helpful to support the overall view of an investigational product. To assess risk profiles for new technologies stratified on its FDA approval or clearance pathway, the number and types of adverse and serious adverse events were extracted from registered clinical trials submitted in the device application.

Results

Between 2014 and 2024, 790 medical devices were FDA-approved or cleared for pulmonary, sleep, or critical care medicine indications. Of the new technologies, 98% (774 of 790) were FDA-cleared against existing technologies. Of these, 0.8% (6 of 790) and 1% (10 of 790) of the devices were FDA-approved through the PMA and De Novo pathways, respectively.

New Technologies Based on Intended Use

A summary of devices based on their intended use and availability of clinical data is reported in Table 1. Of the devices, 23% (178 of 790) were indicated to deliver oxygen or another substrate, 23% (180 of 790) were indicated for monitoring the patient’s condition, and 11% (87 of 790) were for interventional purposes. Most (50% or 5 of 10) De Novo devices were for monitoring, whereas most (67% or 4 of 6) PMA devices were for interventional or therapeutic use.

Table 1.

Number of New FDA-Approved or FDA-Cleared Technologies in Pulmonary/Sleep/Critical Care Medicine Between 2014 and 2024, Stratified by Subspecialty and Availability of Clinical Data Evaluating Intended Use

Subspecialty	510 (k)		PMA		De Novo
	No. of Technologies	Technologies With Clinical Data	No. of Technologies	Technologies With Clinical Data	No. of Technologies	Technologies With Clinical Data
Generala	344	3	0	NA	1	0
Delivery	173	2	2	1	3	2
Monitoring	175	25	0	NA	5	2
Intervention	82	7	4	4	1	1
Total	774	37	6	5	10	5

FDA = US Food and Drug Administration; NA = not applicable; PMA = Premarket Approval.

^aRefers to all devices which do not deliver therapeutics (including gaseous delivery), provide monitoring support, or perform interventional care.

Clinical Trial Data Supporting New Technologies

A total of 95% (737 of 774) of 510(k) devices did not require clinical data to achieve FDA clearance. A total of 50% (5 of 10) of De Novo devices and 83% (5 of 6) of PMA devices had clinical evidence to support their favorable regulatory review. The quality of evidence varied according to the level of evidence required by regulators for each approval pathway. A total of 100% (4 of 4) of intervention devices that were FDA-approved using the PMA pathway had clinical trial data published on clinicaltrials.gov that followed CONSORT guidelines (e-Table 1, Table 2). Study populations and eligibility criteria were described in approval documents for all trials. The primary end point for all trials was an objective disease biomarker, and no trials relied on patient-reported outcome measures (PROMs) as primary end points. When complete clinical data were provided, 20% (1 of 5) of trials for De Novo devices were blinded to investigators (over-the-counter device to assess risk of sleep apnea: No. DEN230041) along with 20% (1 of 5) of trials for PMA devices (Zephyr Endobronchial Valve System: No. P180002). A total of 40% (2 of 5) of trials for PMA devices were not blinded to investigators (Remede System: No. P160039 and Spiration Valve System: No. P18007), and all other trials for De Novo or PMA devices did not describe the blinding method (80% [4 of 5] and 40% [2 of 5], respectively). The statistical plan to analyze the primary end point was described for 24% (9 of 37) of 510(k) devices, 20% (2 of 10) of De Novo devices, and 83% (5 of 6) of PMA devices. The median number of trials for devices with linked clinical trial data was 1, with the number of participants in each study ranging from 24 to 1,325. Of the devices, 80% (4 of 5) of PMA device studies explicitly involved randomization compared with 60% (3 of 5) of De Novo device studies.

Table 2.

New Medical Technologies in Pulmonary/Sleep/Critical Care Medicine Between 2014 and 2024 With Published Linked Clinical Trial Data (clinicaltrials.gov) Included in US FDA Summary/Decision Summary^a

510 (k) Devices
Device (Sponsor; Year; FDA Submission No.)	Study Design	Participants	Interventions	Outcomes	No.	Randomized	Blinding	Statistical Plan to Test Primary End Point
ANNE Sleep (Sibel; 2022; K220095)	Prospective, multicenter, cohort	Adults with moderate to severe OSA	Diagnostic performance of a wireless dual-sensor system (ANNE sleep) compared with reference standard polysomnography	Apnea-Hypopnea Index	225	NA	Blinded	Sensitivity, specificity, accuracy, area under the curve of the receiver operating characteristic curve
NightOwl (Ectosense MV, 2022; K220028)	Prospective, multicentric, cohort	Adults suspected of having sleep apnea	All patients underwent simultaneous polysomnography and peripheral arterial tonometry with home sleep apnea test	Agreement between peripheral arterial tonometry with home sleep apnea testand polysomnography to the interrater agreement of polysomnography	167	NA	Blinded	Two proportion z test
AcuPebble SA (Acurable; 2021; K210480)	Prospective, observational, cohort	Patients referred for evaluation of sleep apnea	Wearable sleep diagnosis device compared with polysomnography	Evaluate the sensitivities and specificities for diagnosis of sleep apnea of a wearable sleep diagnosis technology vs existing gold standard	150	None	Blinded	Likelihood ratios
Pivot Breath Sensor (Carrot; 2021; K201206)	Prospective, observational, cohort	US adults who smoke	Use of an exhaled carbon monoxide breath monitor	Attitudes toward quitting smoking, smoking behavior, quit attempts, and use experience	234	No	No data	t test and χ2 analysis
The iNAP One Sleep Therapy System (Somnice; 2020; K193460)	Prospective, self-controlled, first-night order, cross over	Adults with moderate to severe OSA	Evaluate the efficacy and safety of the iNAP sleep therapy system in adults with OSA	Apnea-Hypopnea Index	32	NA	Blinded	Not reported
DROWZLE (Resonea; 2019; K173974)	Prospective clinical performance	Adult patients referred for sleep study with suspected sleep apnea	Normal and disturbed breathing may be detected by a consumer smartphone without physical connections to the patient using novel algorithms to analyze ambient sound	Ability of acoustic algorithms to detect apnea-hypopnea index < 15 vs > 15 on polysomnography	91	NA	NA	t test and χ2 analysis
Astral 100/150 (ResMed; 2018; K172875)	Prospective, randomized, cross over	Adults with respiratory failure	Compare the automatic expiratory positive airway pressure algorithm with a fixed manual expiratory positive airway pressure in iVAPS (Intelligent Volume-Assured Pressure Support) mode on an Astral mixed mode ventilator	Oxygen Desaturation Index 4%	38	Yes	Blinded	Noninferiority
VPAP Adapt SV VPAP Tx S9 VPAP Tx (ResMed; 2016; K161487)	Multicenter, randomized, parallel cross over	Adult patients with heart failure and reduced ejection fraction	Increased risk of all-cause and cardiovascular mortality in patients with heart failure with reduced ejection fraction and predominant central sleep apnea, randomized to adaptive servo ventilation vs the control group	Optimal medical therapy for heart failure alone, or in combination with adaptive servo-ventilation	1325	Blinded	Investigators	Cox hazard model
Cloud9 System (inSleep Technologies; 2015; K150365)	Prospective, multicenter, reversal cross over	Adult habitual snorers without sleep apnea	CPAP titration study at 2, 4, or 6 cm H2O (night 2) to examine snoring responses to step increases in nasal pressure, a treatment night at optimal pressure (night 3), followed by baseline night (night 4)	Snoring frequency was quantified as a percentage of sleep breaths at thresholds	24	NA	Not reported	Mixed model linear regression

PMA Devices
Device (Sponsor; Year; FDA Submission No.)	Study Design	Participants	Interventions	Outcomes	No.	Randomized	Blinding	Statistical Method
Et Control (Datex-Ohmeda; 2022; P210018)	Prospective, randomized, noninferior	Adults scheduled to undergo general inhaled anesthesia	Use and initiate Et Control after the airway is secured and mechanical ventilation is initiated	Percent duration without large deviation of EtAA based on extraction algorithm, EtAA based on clinician's or investigators' recorded target values of anesthetic agent and oxygen, EtO2 based on extraction algorithm, EtO2 based on clinician's or investigators' recorded target values of anesthetic agent and oxygen	220	Yes	Not reported	Using t test, comparison of the weighted average percent duration
Remede System (RESPICARDIA; 2017; P160039)	Prospective, multicenter, randomized	Adults with moderate to severe central sleep apnea	Device implant, optimal medical therapy, and device initiation 1 mo postimplant	Proportion of participants experiencing a reduction in apnea-hypopnea index and freedom from related serious adverse events within 12 mo	151	Yes	No	z statistic for comparison of means
Zephyr Endobronchial Valve System (Pulmonx; 2018; P180002)	Prospective, multicenter, randomized	Adults with heterogenous emphysema	Assess the safety and efficacy of the EBV and procedure (with pulmonary rehabilitation) compared with optimal medical management (with pulmonary rehabilitation) in patients with heterogeneous emphysema	Percentage of study patients in the Zephyr EBV treatment arm who met the threshold of ≥ 15% improvement in postbronchodilator FEV1 as compared with the control arm at 1 y	190	Yes	Yes	z statistic for comparison of means
Spiration Valve System (Gyrus ACMI; 2018; P18007)	Multicenter, open-label, randomized	Adults with severe heterogenous emphysema	Evaluate the effectiveness and safety of the Spiration Valve System vs optimal medical management	Difference in mean FEV1 from baseline to 6 mo	172	Yes	Open label	95% Bayesian credible intervals for the difference between treatment and control arms

EBV = Emphasys Endobronchial Valve; EtAA = end-tidal anesthetic agent; EtO2 = end-tidal oxygen; FDA = US Food and Drug Administration; NA = not applicable; PMA = Premarket Approval.

^aData derived from FDA Summary/Decision Summary information from devices@fda.gov and/or from completed clinical trials associated with the device on clinialtrials.gov. If device has clinical trial results information from both the FDA information and clinicaltrials.gov, data from clinicaltrials.gov was included. Otherwise, data were derived from the FDA Summary/Decision Summary. In approvals with ≥ 2 studies linked in clinicaltrials.gov with a specific FDA number, study information is divided by a /. All included studies follow the CONSORT criteria for clinical trials.

Adverse Events Associated With New Technologies

Of the devices, 83% (5 of 6) of the PMA devices, 40% (4 of 10) of the De Novo devices, and 1.9% (15 of 774) of the 510(k) devices reported adverse events (Table 3). The nature of the adverse events ranged from skin scrapes, chest tightness, or asthma exacerbation for 510(k) technologies, to device equipment malfunction, pneumothorax, and heart failure for PMA devices. Inherent to the FDA definition of class III devices that have the highest level of risk, 7 device-related deaths occurred for PMA devices that were subsequently approved. Consistent with risk stratification in choosing an appropriate regulatory pathway, no device-related deaths occurred in 510(k) or De Novo device clinical trials.

Table 3.

Overview of Reported AEs and SAEs Across New Pulmonary/Sleep/Critical Care Technologies in the United States Between 2014 and 2024 in Published Clinical Trials and/or Data Included in US FDA Summary/Decision Summary and/or with Linked Clinical Studies With Results on clinicaltrials.gov^a

De Novo Technologies (Low or Medium Risk to Patients)
De Novo Devices	Total No. of Patients in Device Trials	No. of Patients Experiencing at Least 1 AE	Total No. of AEs During Trials	Total No. of SAEs During Trials	Deaths Related to Device	Deaths Unrelated to Device	Most Frequent SAEs
Over-the-counter device to assess risk of sleep apnea	620	0	0	0	0	0	NA
Curve Positive Airway Pressure System	36	Unknown	11	0	0	0	Difficulty breathing through study device
Precision Flow HVNI	204	0	Unknown	0	0	0	Unknown
CNEP Airway	54	Unknown	26	0	0	0	Cutaneous erythema

Premarket Approval Technologies (High Risk to Patients)
Premarket Approval Devices	Total No. of Patients in Device Trials	No. of Patients Experiencing at Least 1 AE	Total No. of AEs During Trials	Total No. of SAE During Trials	Deaths Related to Device	Deaths Unrelated to Device	Most Frequent SAEs
Et Control	220	Unknown	81	2	0	0	Elevated systolic blood pressure
NeuRX Diaphragm Pacing System	53	38	165	23	0	4	Device equipment malfunctions
Remede® System	151	74	53	74	0	7	Heart failure
Zephyr Endobronchial Valve System	190	Not reported	857	201	5	0	Pneumothorax
Spiration Valve System	172	Not reported	Not reported	Not reported	2	12	Pneumothorax

510 (k) Technologies (Substantially Equivalent)
510 (k) Device	Total No. of Patients in Device Trials	No. of Patients Experiencing at Least 1 AE	Total No. of AEs During Trials	Total No. of SAEs During Trials	Deaths Related to Device	Deaths Unrelated to Device	Most Frequent SAEs
Winx Sleep Therapy System	146	2	Unknown	2	0	0	Oral tissue discomfort or irritation
Reprocessed RD SET Adt Pulse Oximeter Sensor	12	0	0	0	0	0	NA
ANNE Sleep System	225	0	0	0	0	0	Unknown
NightOwl	167	0	0	0	0	0	NA
AcuPebble SA	150	0	0	0	0	0	NA
Pivot Breath Sensor	234	0	0	0	0	0	NA
Somnera Sleep System	52	10	19	0	0	0	Difficulty breathing
The iNAP One Sleep Therapy System	32	2	2	0	0	0	Chest tightness; waist soreness
Masimo Rad-97 and Accessories	20	1	1	0	0	0	Skin scrape
SomnaPatch	190	0	0	0	0	0	NA
Astral 100/150	42	1	1	0	0	0	Asthma exacerbation
Carbon Monoxide Breath Sensor System (COBSS)	70	0	0	0	0	0	NA
Juno VPAP ST-A	25	4	4	1	0	0	Aerophagia
VPAP Adapt SV VPAP Tx S9 VPAP Tx	1,325	947	Not reported	947	Unknown	Unknown	Cardiac death
DeVilbiss Intellipap2/DeVilbiss BLUE	28	0	0	0	0	0	NA

For the de novo devices, 4 reported clinical data in their FDA submissions; data for 1 device (Retrograde Intubation Set) was not made available in the public domain. AE = adverse event; FDA = US Food and Drug Administration; NA = not applicable; SAE = serious adverse event.

^aData derived from FDA Summary/Decision Summary information at devices@fda.com and/or linked studies on clinicaltrials.gov. In cases where there were multiple studies in the FDA information, only the linked data from clinicaltrials.gov are included.

Discussion

We systematically synthesized the clinical evidence base for the 790 new FDA-approved/FDA-cleared medical devices in chest medicine over the last decade. Most new devices were FDA-cleared using the 510(k) pathway, which establishes substantial equivalence to existing devices without requiring clinical data through the 510(k) regulatory pathway. Most clinical trial data were reserved for high-risk PMA devices or De Novo devices that were low-to-moderate risk but lacked a predicate device for comparison. Most devices with clinical data required by regulators were indicated for interventional or patient monitoring. In understanding FDA thinking that anchors on risk assessment, there are unmet needs that are deferred to other stakeholders beyond regulators. We therefore report a framework for clinical investigators, professional societies, patient advocacy groups, and industry to collaboratively advance technology development in concert with the US regulatory framework for patients with chest medicine conditions (Fig 1).

Figure 1.

Integrated all-stakeholder framework to develop new technologies for physicians, patient advocacy groups, industry partners, and payers within the US regulatory environment.

Some patients died in clinical trials evaluating devices that were ultimately approved through the PMA pathway, reflecting the needs for clinical experience and evidence to support their use to help patients who are critically ill and need life-saving measures.⁹ Such evidence is gained through published reports on appropriate clinical use and management of complications by clinical experts who are the appropriate early adopters, generally through case reports and case series and sometimes through investigator-initiated clinical trials. The Manufacturer and User Facility Device Experience Database maintained by the FDA provides a mechanism for reporting and learning about adverse events with new technologies, and is a critical mechanism for postmarketing surveillance by regulators and clinicians alike.¹⁵

Clinical trials should generally aim to adhere to CONSORT guidelines that describe the conduct of high-quality clinical trials.¹⁶ Device trials face technical nuance in using devices, and blinding may not be possible. Randomization may be unethical without an appropriate sham control. Achieving sufficient power may be challenging, necessitating large-scale registries supported by professional societies to support clinical use and ultimately payer coverage.¹⁷ In contrast to chronic illness, objective improvement may be more relevant at present to critical care devices than available PROMs.¹⁸ However, it is important to recognize that the FDA has embraced the concept of aligning labeling with the global patient experience associated with changes in objective biomarkers that reflect both the underlying pathophysiology and mechanism of any investigational interventions.¹⁹ This regulatory guidance intends to shift focus away from objective biomarkers that have little established real relevance to human health. Efforts to develop valid and responsive PROMs are underway for pulmonary hypertension,²⁰ asthma,²¹ and sleep apnea²² both for clinical trials and practice. An updated list of PROMs is maintained by the American Thoracic Society (https://www.thoracic.org/members/assemblies/assemblies/bshsr/patient-outcome/).

To our knowledge, this is the first scoping review to document the breadth and scope of clinical data to support US FDA approval/clearance of new medical devices relevant to chest medicine in recent years, resulting in a framework to support an innovation ecosystem over the next decade (Fig 1). From a clinician standpoint, patients in an ICU setting face a different set of risks (and probable benefits compared with probable risks) compared with outpatients with chronic health conditions, highlighting the importance of our study that is specific to chest medicine.⁶ Similar evaluations have been conducted over the prior 2 decades in other clinical fields, such as gastroenterology,²³ cardiology,²⁴ and orthopedics,²⁵ and demonstrate consistency in how FDA applies a standardized risk-based framework for new medical devices across fields.⁹ A total of 98% of chest medicine devices used the 510(k) pathway to achieve FDA clearance by demonstrating substantial equivalence to an established device. Beyond chest medicine, this pathway has come under scrutiny: some of the predicate devices used in 510(k) submissions have been recalled or are out-of-date with current practice, and 8% of predicate devices have ever had clinical data evaluated by FDA in one analysis.²⁶ To address this scrutiny, FDA released guidance documents that emphasize use of contemporary predicate devices and generating clinical data to support new indications for use, differences between investigational and predicate devices, and historical risks associated with predicate devices.²⁷

Regulators in different regions have varying perspectives on risk assessment that shift over time. In Europe, for example, formal working groups of clinical investigators are advocating for changes in clinical trial expectations and reporting for clinical trials to align the bar to achieve a Conformité Européenne mark to be more stringent.²⁸ One impact in this approach is that the cost and timeline for device development could expand. On one hand, this potentially limits the pipeline for new devices, with an intent that the resultant pipeline would be of greater quality in terms of a higher bar for patient safety. Enhancing clinical trial requirements might narrow the labeling of new products because rigorous clinical trials may need to be conducted for each labeled indication.⁹ In the United States, separate efforts are underway to develop pragmatic trial designs that follow established pathways to FDA approval/clearance, such as single-arm studies designed to provide real-world evidence on routine use. One example is the recent Transbronchial Biopsy Assisted by Robot Guidance in the Evaluation of Tumors of the Lung (TARGET) trial, which evaluated the use of robotic-associated bronchoscopy in sampling indeterminate peripheral pulmonary lesions.²⁹

Limitations of our study are several. Not all clinical data used by regulators are publicly available; therefore, we are able to report on clinical evidence that is available to providers and patients. We focus our analysis on rigor in public descriptions of primary end points used to evaluate efficacy by industry sponsors and FDA, recognizing that secondary and exploratory end points may provide additional data that can impact labeling. FDA allows sponsors to include human data beyond clinical trials alone, such as animal models, human factors testing, or evidence from the peer-reviewed literature.^30,31 We focus on original applications rather than modifications to devices that were originally approved before 2014, so that our analysis focuses on FDA’s current thinking, highlighting gaps and opportunities for partnership among academia, professional societies, industry sponsors, and clinicians. For example, FDA allows postmarketing data to support label expansion after initial approval.³¹

Interpretation

We evaluated all FDA-regulated pulmonary, sleep, and critical care medicine technologies that comprise chest medicine and came to market in the last decade. The innovation pipeline supporting the future of chest medicine is robust. Action items include anticipating gaps in clinical evidence for disruptive 510(k) devices and pragmatic studies supporting appropriate use of PMA devices. Supporting these recommendations will facilitate an environment of continued innovation in the coming decade.

Funding/Support

E. D. S. is supported by the NIH [Grant 1K23DK134752].

Financial/Nonfinancial Disclosures

The authors have reported to CHEST the following: E. D. S. consulted for Ardelyx, Cook, Laborie, Mahana, Mylan, Neuraxis, Phathom, Salix, Regeneron, Sanofi, and Takeda and has a patent with the Regents of the University of Michigan. None declared (J. T. G., J. A. B.).