AI & Reanalysis: Close the gap on your past open cases

Introduction

Reanalysis. With demand for testing accelerating, this is quickly becoming a painful topic for labs. Every patient matters, and deserves the latest science can offer. But reanalyzing an ever-growing number of cases is unsustainable in the long run. If 50% of your cases are unsolved, and you revisit them every 6 or 12 months, 10,000 will quickly turn into 20,000, 30,000 …100,000. At some point, the number of open cases will exceed the manpower available for backtesting. Add to that challenge the 1.5 million new articles added to PubMed every year, keeping up with new information and matching it to cases months or years old becomes a virtually impossible task – for humans.

Why reanalyze?

The literature takes a positive stand on the value of reanalysis. Wegner et al. reanalyzed the exome of 40 patients with a possible Mendelian disorder. Published in Genetics in Medicine in 2017, they were able to reach a definitive diagnosis in 10% of cases reanalyzed. At the time of the original diagnosis, the literature was “weak, nonexistent, or not readily located”. In a more recent publication, Genetics in Medicine October 2018, Basel Salmon et al. reanalyzed 84 probands – the reanalysis was performed on the Emedgene platform. In total, new diagnoses were established in 13/84 individuals (15.5%). Reasons the diagnoses weren’t reached originally were “incorrect interpretation of the clinical context and absence of OMIM entry”. Previously unknown gene-disease associations were discovered in three of the probands.

Wegner et al. also estimated that approximately 250 gene-disease (OMIM) and 9,200 variant-disease associations (HGMD) are reported annually. The actual number of discoveries is much higher, as it takes time between an initial discovery and publication, validation in additional studies, and inclusion in a database. This fast-growing body of knowledge means patient care will be best served by reanalysis with new data.

But, this is a business

However, this becomes challenging when one factors in costs. What is the business case for reanalysis? How often is it performed? Who foots the bill? What is the time, effort and cost taking into consideration the growing number of unsolved cases?

Yes, reanalysis has a CMS code, and revisiting the occasional case – by a clinician’s request, for example – is simple.

Many labs reanalyze past cases periodically, say every 6 or 12 months, and have found a way to do so within their business model. Whatever your approach is now, it’s not sustainable.  The costs and time involved with mass reanalysis will overwhelm lab productivity and profitability.

Automated reanalysis is the answer

Incorporating automated reanalysis is the key to handling the growth in genomics knowledge to better serve patients, and doing it profitably.

Automated reanalysis uses AI to “read” new information and understand if it applies to existing open cases. When you do reanalyze, you can focus your efforts only on the percentage of cases that you may be able to solve now. This changes the economics of backtesting for your lab, keeps productivity from being impacted, and increases lab yield.


Sources
1 Systematic reanalysis of clinical exome data yields additional diagnoses: implications for providers,
Aaron M. Wenger PhD, Harendra Guturu PhD, Jonathan A. Bernstein MD, PhD & Gill Bejerano PhD
Genetics in Medicine volume 19, pages 209–214 (2017)
2 Improved diagnostics by exome sequencing following raw data reevaluation by clinical geneticists involved in the medical care of the individuals tested, Lina Basel Salmon MD, PhD, Naama Orenstein MD, Keren Markus-Bustani MSc, Noa Ruhrman-Shahar MD, Yael Kilim MSc, Nurit Magal PhD, Monika Weisz Hubshman MD, PhD & Lily Bazak PhD, Genetics in Medicine (2018)

Learn more about Reanalysis with Emedgene's AI

Recent Posts

Learn More About Reanalysis with Emedgene's AI
2018-12-06T14:35:25+00:00