Germline-focused analysis of tumour-detected variants in 49,264 cancer patients: ESMO Precision Medicine Working Group recommendations
Authors: Z. Kuzbari, C. Bandlamudi, C. Loveday, M.F. Berger, D. Mandelker, C. Turnbull
Show all authors
- Strategic filtering improves the germline conversion rate with minimal loss of true germline variants present in the tumour.
- Germline conversion rate of filtered tumour-detected variants is very high (>80%) for genes such as BRCA1, BRCA2 and PALB2.
- Germline conversion rate of filtered tumour-detected variants is very low (<2%) for genes such as APC, TP53 and STK11.
- Germline follow-up should involve multidisciplinary expertise and follow expert guidance regarding tumour context.
- To our knowledge, this is the largest germline focused-analysis to date, including 49,264 paired tumour-normal samples.
The ESMO Precision Medicine Working Group was reconvened to update its 2018/19 recommendations on follow-up of putative germline variants detected on tumour-only sequencing, which were based on an analysis of 17,152 cancers.
We analysed an expanded dataset including 49,264 paired tumour-normal samples. We applied filters to tumour-detected variants based on variant allele frequency (VAF), predicted pathogenicity and population variant frequency. For 58 cancer susceptibility genes, we then examined the proportion of filtered tumour-detected variants of true germline origin (germline conversion rate (GCR)). We conducted sub-analyses based on the age of cancer diagnosis, specific tumour types and “on-tumour” status (established tumour-gene association).
Analysis of 45,472 non-hypermutated solid malignancy tumour samples yielded 21,351 filtered tumour-detected variants of which 3,515 were of true germline origin. 3.1% of true germline pathogenic variants were absent from the filtered tumour-detected variants. For genes such as BRCA1, BRCA2, and PALB2, the GCR in filtered tumour-detected variants was >80%; conversely for TP53, APC and STK11 this GCR was <2%.
Strategic germline-focused analysis can prioritise a subset of tumour-detected variants for which germline follow-up will produce the highest yield of most actionable true germline variants. We present updated recommendations around germline follow-up of tumour-only sequencing including (i) revision to 5% for the minimum per-gene GCR, (ii) inclusion of actionable intermediate penetrance genes ATM and CHEK2 (iii) definition of a set of seven ‘most-actionable’ cancer susceptibility genes (BRCA1, BRCA2, PALB2, MLH1, MSH2, MSH6, RET) in which germline follow-up is recommended regardless of tumour type.