Skip to main content
The NCI Community Hub will be retiring in May 2024. For more information please visit the NCIHub Retirement Page:https://ncihub.cancer.gov/groups/ncihubshutdown/overview
close
eeDAP studies Logo
  • Discoverability Visible
  • Join Policy Restricted
  • Created 13 Jan 2017

eeDAP studies

►

Wiki

Device Advice for AI and Machine Learning Algorithms In Medical Imaging

Version 47
by (unknown)
Version 48
by (unknown)

Deletions or items before changed

Additions or items after changed

1 Here is a document of links pointing to FDA guidance documents and other forms of communication. It was compiled by googling CDRH regulatory buzz words. We hope this document is useful. You might also refer to [https://www.jpathinformatics.org/viewimage.asp?img=JPatholInform_2020_11_1_22_291538_f1.jpg Figure 2] of this paper: Marble et al. (2020). A regulatory science initiative ..." J Pathol Info, 11(1), 22. https://doi.org/10.4103/jpi.jpi_27_20
2
3 Caveat to all: You won't know what's required for your submission until you ask in a Q-sub.
4
5 === Pre-submission meeting ===
6
7 In CDRH, the best advice is to know about pre-submission meetings.
8 * [https://www.fda.gov/media/93740/download Slides: Pre-Submissions and Meetings with FDA Staff]
9 * [https://www.fda.gov/media/114034/download Guidance document: Requests for Feedback on Medical Device Submissions: The Pre-Submission Program and Meetings with Food and Drug Administration Staff]
10
11
12
13 === IFU: Indications for Use ===
14 Part of the definition of your medical device are the indications for use.
15 * [https://www.fda.gov/regulatory-information/search-fda-guidance-documents/device-labeling-guidance-g91-1-blue-book-memo Webpage: Indications For Use.]
16
17 It's never to early to start thinking about, researching, and crafting an IFU for your device.
18
19
20
21
22 === Comprehensive Regulatory Assistance ===
23
24 * [https://www.fda.gov/medicaldevices/deviceregulationandguidance/ Webpage: Comprehensive Regulatory Assistance]
25
26
27
28
29 === 510k program ===
30 * [https://www.fda.gov/downloads/MedicalDevices/DeviceRegulationandGuidance/GuidanceDocuments/UCM284443.pdf Guidance Document: The 510(k) Program: Evaluating Substantial Equivalence in Premarket Notifications 510(k)]
31 * [http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfPMN/pmn.cfm 510(k) Premarket Notification Database]
32
33 A Premarket Notification, 510(k), is a submission made to FDA to demonstrate that the device to be marketed is safe and effective by proving substantial equivalence (SE) to a legally marketed device (predicate device).
34
35 * [https://www.fda.gov/regulatory-information/search-fda-guidance-documents/deciding-when-submit-510k-change-existing-device Guidance Document: Deciding When to Submit a 510(k) for a change to an Existing Device.]
36 * [https://www.fda.gov/regulatory-information/search-fda-guidance-documents/deciding-when-submit-510k-software-change-existing-device Guidance Document: Deciding When to Submit a 510(k) for a Software Change to an Existing Device.]
37
38
39
40
41 === de Novos ===
42
43 * Special controls accompany all de Novo classifications.
44
45 * Special controls outline submission requirements for medical devices with similar IFU.
46
47 * Special controls are defined for a device type that can be broad or narrow.
48
49 * [https://www.fda.gov/medicaldevices/deviceregulationandguidance/overview/generalandspecialcontrols/default.htm#special Webpage: Special Controls]
50
51 * [https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfPMN/denovo.cfm De Novo Database]
52
53 If you can't find a de Novo or existing 510k that is a perfect predicate for your device type: Then find and learn about the closest one or two. It's a way to start.
54
55
56
57
58 === PMA ===
59
60 Pursuing a pre-market authorization ('''PMA''') is a regulatory pathway for medical devices that have a higher risk profile than a 510(k). PMAs are not covered here but information can be found by searching for "FDA CDRH PMA" from any internet search engine.
61
62
63 === Software as a Medical Device (SAMD) ===
64
65 * [https://www.fda.gov/medical-devices/digital-health-center-excellence Digital Health Center of Excellence]
66 * [https://www.fda.gov/medical-devices/digital-health/software-medical-device-samd Software as a medical device]
67 * [https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-software-medical-device AI/ML in SAMD]
68 * [https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-aiml-enabled-medical-devices List of AI/ML enabled devices]
69
70 * Computer aided detection (CADe) guidance - Radiology'''
71 * [https://www.fda.gov/regulatory-information/search-fda-guidance-documents/computer-assisted-detection-devices-applied-radiology-images-and-radiology-device-data-premarket Webpage: Non-clinical=Stand-alone=No human in the loop:]
72 * [https://www.fda.gov/regulatory-information/search-fda-guidance-documents/clinical-performance-assessment-considerations-computer-assisted-detection-devices-applied-radiology Webpage: Clinical=Reader Study=Human in the loop:]
73
74 * [https://www.fda.gov/regulatory-information/search-fda-guidance-documents/shelf-software-use-medical-devices Guidance on Off-The-Shelf Software Use in Medical Devices]
75
76
77
78 === Quantitative Imaging ===
79 [https://www.fda.gov/regulatory-information/search-fda-guidance-documents/technical-performance-assessment-quantitative-imaging-device-premarket-submissions Guidance Technical Performance Assessment of Quantitative Imaging in Device Premarket Submissions]
80
81
82
83 === Adaptive Algorithms, continuous learning ===
84 * [https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-software-medical-device Webpage, Artificial Intelligence and Machine Learning in Software as a Medical Device]
85 * [https://www.fda.gov/media/122535/download White paper, Proposed Regulatory Framework for Modifications to Artificial Intelligence/Machine Learning (AI/ML)-Based Software as a Medical Device (SaMD)]
86
87
88
89 === Digital Pathology ===
90 * Device type Whole Slide Imaging System
91 * Philips !IntelliSite Pathology Solution (DEN160056: https://www.accessdata.fda.gov/cdrh_docs/reviews/DEN160056.pdf).
92 * Classification Product Code: PSY
93 * Record includes special controls
94 +
* Device type Whole Slide Imaging System
95 +
* [https://www.fda.gov/regulatory-information/search-fda-guidance-documents/technical-performance-assessment-digital-pathology-whole-slide-imaging-devices Guidance for industry and FDA staff - technical performance assessment of digital pathology whole slide imaging devices]
96 +
97
98 === CAD Examples ===
99 * Device type: Medical image analyzer (including breast cancer detection)
100 * Secondlook (P010038: http://www.accessdata.fda.gov/cdrh_docs/pdf/P010038B.pdf)
101 * Originally a PMA. Downclassified to 510k on 01/22/2020 ([https://www.federalregister.gov/documents/2020/01/22/2020-00494/radiology-devices-reclassification-of-medical-image-analyzers FR LINK])
102 * Classification Product Code: MYN
103
104
105
106 * Device type: Radiological computer-assisted diagnostic (CADx) software for lesions suspicious for cancer.
107 * !QuantX (DEN170022: https://www.accessdata.fda.gov/cdrh_docs/reviews/DEN170022.pdf)
108 * Classification Product Code: POK
109
110
111
112
113 * Device type: Radiological Computer Assisted Detection and Diagnosis Software" (CADe ''+'' CADx).
114 * Imagen !OsteoDetect (DEN180005: https://www.accessdata.fda.gov/cdrh_docs/pdf18/DEN180005.pdf)
115 * Classification Product Code: QBS
116
117
118
119
120 * Device type: Radiological computer aided triage and notification software.
121 * Viz.AI !ContaCT (DEN170073: https://www.accessdata.fda.gov/cdrh_docs/reviews/DEN170073.pdf)
122 * Classification Product Code: QAS
123
124
125
126
127 * Device type: Retinal diagnostic software device
128 * IDx IDx-DR (DEN180001: https://www.accessdata.fda.gov/cdrh_docs/reviews/DEN180001.pdf)
129 * Classification Product Code: PIB
130
131
132
133 * Device type: Image acquisition and/or optimization guided by artificial intelligence
134 * Bay Labs Caption Guidance (DEN190040: http://www.accessdata.fda.gov/cdrh_docs/pdf19/DEN190040.pdf)
135 * Classification Product Code: QJU
136
137
138
139 * Device type: Software algorithm device to assist users in digital pathology
140 * Paige.AI Paige Prostate (DEN200080: https://www.accessdata.fda.gov/cdrh_docs/pdf20/DEN200080.pdf)
141 * Classification Product Code: QPN
142 * Presentation and discussion about the Paige.AI decision summary convened by the Pathology Innovation Collaborative Community (June 2022): [https://pathologyinnovationcc.org/presentations/jun-2022-decision-summary LINK]
143
144 === Presentations in this space ===
145
146 * "Tutorial on Reader Study Designs and MRMC Analysis"
147 * FDA internal training, April 8, 2022
148 * Pathology Innovation Collaborative Community Webinar, August 5, 2022
149 * Brandon Gallas, Research Mathematical Statistician, Division of Imaging, Diagnostics, and Software Reliability OSEL, CDRH, FDA
150 * [[File(20220805-PIcc-MRMCstudyDesigns-Gallas.pdf)]]
151 * [https://vimeo.com/751682643 Video MRMC tutorial]: 1 hour 20 seconds
152
153 * "ROC curves: Receiver Operating Characteristic Curves"
154 * FDA internal training, April 8, 2022
155 * Pathology Innovation Collaborative Community Webinar, August 5, 2022
156 * Brandon Gallas, Research Mathematical Statistician, Division of Imaging, Diagnostics, and Software Reliability OSEL, CDRH, FDA
157 * [[File(20220805-PIcc-ShortROCtutorial.pdf)]]
158 * [https://vimeo.com/751670299 Video ROC tutorial]: 6 minutes 17 seconds
159
160 * "Regulatory Considerations and Assessment of AI/ML Devices"
161 * Yale, January 29, 2020, New Haven, Connecticut
162 * Nicholas Petrick, Deputy Division Director, Division of Imaging, Diagnostics, and Software Reliability OSEL, CDRH, FDA
163 * [[File(Petrick_RegulatoryConsiderationsAssessmentAI2020-01-29.pdf)]]
164
165 * "AI and Digital Pathology: Regulatory Perspective"
166 * Pathology Visions, October 8, 2019, Orlando, FL.
167 * Shyam Kalavar, Senior Scientific Reviewer, Division of Molecular Genetics and Pathology
168 * [[File(S.Kalavar.PathVision2019.pdf)]]
169
170 * "Evaluating Artificial Intelligence Devices at the FDA and Related Collaborations and Initiatives"
171 * ACR Informatics Summit, October 5-6, 2019, Washington, DC.
172 *
173 * Brandon Gallas, !PhD, Research Mathematical Statistician, Division of Imaging, Diagnostics, and Software Reliability OSEL, CDRH, FDA
174 * [[File(20191005ACRinformaticsSummit_BDG-6-FINAL.pdf)]]
175 *
176 * Jennifer Segui, Lead Medical Device Reviewer, Division of Radiological Health, FDA
177 * [[File(J.A.Segui.ACR.Informatics.2019.Slides.FINAL.pdf)]]
178 *
179
180 * "Evaluation and Regulatory Considerations for AI Methods in Medical Imaging"
181 * Society for Imaging Informatics in Medicine Annual Meeting, June 26, 2019, Aurora, Co
182 * Berkman Sahiner, Senior Scientist, Division of Imaging, Diagnostics, and Software Reliability OSEL, CDRH, FDA
183 * [[File(Sahiner_EvaluationRegulatoryConsiderationsAIinMedicalImaging20200626.pdf)]]
184
185 * "Digital Pathology Regulatory Considerations"
186 * Pathology Informatics Summit 5/9/19
187 * Cheng Cui, Senior Scientific Reviewer, Division of Molecular Genetics and Pathology,
188 * [[File(PathologySummit2019_Pittsburgh_ChengCui_FDA.pdf)]]
189
190 === Medical Device Development Tools (MDDT) program ===
191
192 The FDA's Medical Device Development Tools (MDDT) program is a way for the FDA to qualify tools that medical device sponsors can use in the development and evaluation of medical devices. It is a way for the broader community (academia, health providers, and government scientists, not just industry) can impact the regulatory process.
193 * [https://www.fda.gov/medicaldevices/scienceandresearch/medicaldevicedevelopmenttoolsmddt/ Webpage: FDA page, "Medical Device Development Tools (MDDT)"]
194
195
196 === Catalog of Regulatory Science Tools to Help Assess New Medical Devices ===
197
198 * [https://www.fda.gov/medical-devices/science-and-research-medical-devices/catalog-regulatory-science-tools-help-assess-new-medical-devices Link to catalog]
199
200 === Mock Submission ===
201
202 * [https://mdic.org/event/computational-modeling-simulation/ Webpage: MDICx webinar that includes '''a presentation on mock submissions to FDA/CDRH.]'''
203 * [http://mdic.org/wp-content/uploads/2014/05/CMS-Summit-Myers.pdf Slides: Mock Submissions to FDA/CDRH: History and Lessons Learned:] by Kyle Myers, Director DIDSR/OSEL/CDRH/FDA. This presentation was made as the agency was working with MDIC to pursue a mock submission about, "Augmenting a Clinical Study with Virtual Patient Models."
204 * [https://webcache.googleusercontent.com/search?q=cache:C2Ru6HlIKEAJ:https://mdic.org/project/virtual-patient-vp-model/+&cd=7&hl=en&ct=clnk&gl=us&client=firefox-b-1-d Webpage: All the '''mock submission documents''' to and from the MDIC team and FDA.] Actually, this cached web link seems to now point to the updated Virtual Patient Page at MDIC which is missing the FDA submission feedback.
205 * The mock submission was followed quickly by actual submissions.
206 * Slides, Mock Submissions updated [[File(20190412-HTTMockSubmissions.pdf)]]
207 * [https://mdic.org/project/virtual-patient-vp-model/ Updated Virtual Patient Page at MDIC]
208 * The virtual patients mock submission was preceded by [http://clinchem.aaccjnls.org/content/56/2/165 "Protein-Based Multiplex Assays: Mock Presubmissions to the US Food and Drug Administration"], Regnier et al.
209 * [https://academic.oup.com/clinchem/article/56/2/165/5622492#supplementary-data Supplementary Materials]