October 2020
Alliance Visions 2020
The Alliance held a 1-day meeting on October 30th, 2020 immediately after Visions. It was our first online, and 5th meeting overall. Below is the summary and all content (open and freely available).
The meeting was hosted by MGH, with support from 5 workgroups (see below).
The Alliance Vision 2020 Meeting Agenda
Aims of the Meeting
Provide a status update of the Alliance
Introduce The Alliance Project Manager
Provide additional perspectives and updates from related groups and stakeholders
Enable new members a brief overview of the workgroups
Deliver cutting-edge information regarding Tumor-infiltrating lymphocytes in breast cancer’ plenary talk by Roberto Salgado
Status update Collaborative Community
Four Main Topics
Welcome
10:00-10:10 Welcome - General remarks (Esther + Brandon + Joe)
10:10-10:20 Introduction PM Ula Green (Ula)
Plenary Talk
10:30-10:50 Plenary Session - “Machine Learning and TILs: Quo Vadis?” (Roberto Salgado, Belgium)
10:50-10:55 Discussion
Roberto Salgado is board certified in Anatomical Pathology since 2006. After his graduation as an Anatomical Pathologist he worked, besides as a clinical pathologist, also as a consultant Pathologist for the Breast International Group (BIG), and as a member of the Aurora Working Group helped develop the Aurora Molecular Screening Program for Metastatic Breast Cancer Patients. He was also a member of the Aurora Molecular Advisory Board of the Breast International Group. He’s a scientific collaborator of the Immuno-Task Force of the Breast International Group and an Honorary Research Fellow at the Division of Research at the Peter Mac Callum Cancer Center, Melbourne, Australia. Currently he’s chairing an International Consortium of Pathologists, namely the “International Immuno-Oncology Biomarkers Working Group that develops guidelines for the assessment of Immuno-Oncological Biomarkers in cancer. His strategic views on Oncology have been published in major international high impact factor journals including Nature Reviews Clinical Oncology, Nature Reviews Drug Discovery and Lancet Oncology. He has specific expertise in the Molecular Pathology of solid tumours, validation of biomarkers, innovative clinical trials incorporating genomics and development of biomarkers in immunotherapeutic trials, with a focus on Tumor Infiltrating Lymphocytes (TILs), with publications in Lancet Oncology, Nature, Nature Medicine and Journal of Clinical Oncology.
Summary
Machine Learning and TIL’s – Quo Vadis?
By Roberto Salgado
Dr. Salgado provided an outline of the role of Tumor infiltrating lymphocytes (TIL’s) in breast cancer. He started by providing a definition – followed by novel data in young triple-negative breast. Cancer patients. (These data are not published and are blurred). He also outlined the relevance of machine learning tools in relation to TIL’s. He mentioned a recent commentary that outlines some of the unintended consequences of regulatory decisions. He outlined in concrete terms how assay approval could be improved. As part of this topic he outlined a novel quality control program for testing and validation in anatomic pathology. He made 2 concrete project suggestions to the Alliance:
1) A best practices paper that many have started and the Alliance could contribute to as well.
2) Proposals for additional articles for a Special Issue in Cancers with a specific focus on topics related to the TIL’s workgroup and the Alliance.
Finally, he pointed to two publications outlining Phase I of the TIL’s in breast cancer consortium. The talk was well-received and followed by a discussion how to continue to work together across the Allince and TIL’s WG.
Breakout sessions
10:55-11:00 Overview of the breakout sessions
11:00-12:00 Breakout sessions
1 Pre-analytics
Leads: Amanda Lowe, Matt Leavitt
Build out framework and table of content for guideline regarding pre-analytics
We should consider basic pre-analytical) elements (e.g., position of tissue on slide)
Aging of slides - how do older slides impact pre-analytics ?
Next steps: Meet with the full team to review and designate who will write what
Reaching out to CAP, etc. to harmonize not duplicate
2 Standards
Leads: Markus Herrmann, Mike Isaacs
Working on proposal to HTT group - opportunities to collaborate with respect to standardization of dataset (help truthing with data standardization)
How to convert datasets into standard format
HTT group - compiling and annotating data
Create image data repository in DICOM format - images and annotations
Use case - international COVID19 repository
Waiting for data to come in to standardize (Truthing group to help standardize data)
Collecting microscopy data from around the world - autopsy cases
Data available to researchers
Pursue model - individual institutions can contribute data to WHO - interface which allows pathologist to provide autopsy reporting data
Data sharing internationally between commercial and public entity is challenging - process of collecting data, images and corresponding autopsy
3 High-throughput truthing (HTT) group
Leads: Sarah Dudgeon, Brandon Gallas, Hetal Marble
Need domain expert pathologists to lead training module
Combine efforts to create continuing education
Current response to FDA review of medical device development tool proposal
Key element is defining how patients will benefit from the proposal/tool
Pilot study data compilation
Consider creating subgroups
Plan new submission to FDA by 2021
4 Practical Use Cases/Machine learning
Lead: Matthew Hanna
Pathologists needs practical workflow efficiencies
Importance of Interoperability considerations from industry
Early adopters have either developed their own software/hardware or have collaborated with industry to initiate their digital transformations,
We need standardizations and further maturity in order to scale for everyday workflows.
5 Reimbursement
Leads: Esther Abels, Laura Lasiter, Joe Lennerz
Esther and Joe provided and update (see here)
Current landscape
High cost in short term
Return on investment: important for laboratories to set their budget and or strategy for adoption
Key question: Will digital pathology improved outcomes or just improved operation (e.g. turnaround times). Answer will determine approach to demonstrate clinical utility
Specific use cases to advance reimbursement
White paper will entail “Actionable next steps for payor strategy”
Next steps
Include payors – Need to understand pain points
Educate medical directors of health plans
Population health analysis
Focus on individual diseases (i.e., concrete tangible use cases)
Collaborative trial (joint effort by multiple stakeholders: shared risk, shared benefit)
Collaborative Community
12:00-12:30 “Alliance CC” - towards a Collaborative community
12:30-1:00 Open discussion, breakout rooms, meet and greet. wrap up.
Collaborative Community
In this last part Joe Lennerz provided an update on the progress regarding creation of a collaborative community in the field of digital pathology. In the regulatory space in the US, collaborative communities are -in contrast to the lay-term- specific groups and the FDA provided a toolkit. The presentation provided an status update and concrete tasks for finalizing the current draft charter. Other point mentioned were:
The Collaborative Community will exist separately from the Alliance
Charter drafting sessions (open to public) => see also here
11/10 at 3pm ET
11/13 at 10am ET
Establish a steering committee
Key aim is to establish the collaborative community and start delivering
Exciting approach for the field (“us”) to engage formally with the regulatory (FDA)
Enable various stakeholders to directly provide input into regulatory frameworks
We suggest the following resources as a great starting point to prepare for the drafting sessions:
Study existing communities
Consider assuming a role
Join the drafting sessions and help craft the structure of the community