Should Histology Play a Role in Deciding Which Colorectal Carcinomas Should Be Tested for Lynch Syndrome?
Suntrea TG Hammer, Joel K Greenson. University of Michigan Health System, Ann Arbor, MI
Background: Recent publications have suggested that it is cost effective to test all colorectal cancers (CRCs) in order to identify patients with Lynch syndrome. Such testing typically includes microsatellite instability (MSI) testing, IHC stains for mismatch repair proteins (MMR) and Braf V600E analysis. We have previously published a morphology based predictive model of microsatellite instability (Am J Surg Pathol. 2009 Jan;33(1):126-33.) that we have used to select cases for Lynch syndrome testing. Starting in August of 2011, however we began testing all CRCs for MSI, MMR, and Braf if MSI-H. We reviewed all CRCs tested during the past year to see if we would have missed any Lynch syndrome cases had we used our morphologic predictor to direct testing.
Design: We performed a database review for all colorectal carcinoma resections performed from 8/1/2011 through 7/31/2012 which yielded 88 specimens. We then omitted cases with preoperative chemotherapy/radiation and carcinomas arising in polypectomy specimens (n=14). Each case was reviewed for age, tumor location, histology, MSI status, IHC for MSH2, MSH6, MLH1, PMS2, and Braf mutation analysis. The histologic features used in our predictive model were number of tumor infiltrating lymphocytes/ high power field, Crohn's like reaction, mucinous differentiation, dirty necrosis, and tumor differentiation. These data were entered into our web based predictor of MSI (http://sitemaker.umich.edu/gruber.lab/files/msi_pre.htm) to see how well we could predict MSI status. We also calculated the total generated charge and potential savings if we had used this morphologic model rather than testing all tumors.
Results: We previously used a >5% likelihood of MSI-H as a cut-off to test tumors for Lynch syndrome. Thirty-six of 74 cases had morphologic path scores predicting a >5% likelihood of MSI-H. All eleven cases of MSI-H colorectal carcinoma (including 8 cases of Lynch syndrome) would have been selected for testing by this method. Thirty-eight cases of MSS CRC had a likelihood of MSI below 5%. The charge for MSI and IHC testing these 38 MSS stable CRCs was $111,758 ($2941/case).
Conclusions: Using our histology based MSI predictive model we could have omitted testing 38 of 74 cases of CRC without missing any MSI-H or Lynch tumors for a total savings of $111,758. This would cut the cost of screening for Lynch syndrome and MSI by greater than 50%.
Wednesday, March 6, 2013 9:30 AM
Poster Session V # 97, Wednesday Morning