Clinical impact of extensive molecular profiling in advanced cancer patients

Previous precision medicine studies have investigated conventional molecular techniques and/or limited sets of gene alterations. The aim of this study was to describe the impact of the next-generation sequencing of the largest panel of genes used to date in tumour tissue and blood in the context of institutional molecular screening programmes. DNA analysis was performed by next-generation sequencing using a panel of 426 cancer-related genes and by comparative genomic hybridization from formalin-fixed and paraffin-embedded archived tumour samples when available or from fresh tumour samples. Five hundred sixty-eight patients were enrolled. The median number of prior lines of treatment was 2 (range 0–9). The most common primary tumour types were lung (16.9%), colorectal (14.4%), breast (10.6%), ovarian (10.2%) and sarcoma (10.2%). The median patient age was 63 years (range 19–88). A total of 292 patients (51.4%) presented with at least one actionable genetic alteration. The 20 genes most frequently altered were TP53, CDKN2A, KRAS, PTEN, PI3KCA, RB1, APC, ERBB2, MYC, EGFR, CDKN2B, ARID1A, SMAD4, FGFR1, MDM2, BRAF, ATM, CCNE1, FGFR3 and FRS2. One hundred fifty-nine patients (28%) were included in early phase trials. The treatment was matched with a tumour profile in 86 cases (15%). The two main reasons for non-inclusion were non-progressive disease (31.5%) and general status deterioration (25%). Twenty-eight percent of patients presented with a growth modulation index (time to progression under the early phase trial treatment/time to progression of the previous line of treatment) >1.3. Extensive molecular profiling using high-throughput techniques allows for the identification of actionable mutations in the majority of cases and is associated with substantial clinical benefit in up to one in four patients. Electronic supplementary material The online version of this article (doi:10.1186/s13045-017-0411-5) contains supplementary material, which is available to authorized users.

Previous precision medicine studies have investigated conventional molecular techniques and/or limited sets of gene alterations [1][2][3]. We describe here the impact of the next-generation sequencing of the largest panel of genes used to date in tumour tissue and blood in the context of institutional molecular screening programmes. The eligibility criteria, methods of sequencing and statistics are described in Additional file 1.
Between January 1, 2014, and June 30, 2015, 568 patients were enrolled in the study. Their characteristics are summarized in Additional file 2: Table S1 and Additional file 3: Figure S1.
Two hundred ninety-two (51.4%) patients had at least one genetic alteration that was considered actionable by the molecular tumour board. Molecular profiles by tumour type are presented in Additional file 5: Figure S3.
In the GM group, 65 patients were evaluable for the response to treatment analysis. The disease control rate (objective response rate + stable disease) was 47.7% (Additional file 2: Table S3). The median progressionfree survival (PFS) was 3 months. The median overall survival was 8.5 months (range 5.5-11.5 months).
Thirty-nine patients with coupled primary and metastatic tumours were analysed to evaluate the correlation between the molecular screening results of the two samples. Twenty-six patients (67%) had at least one mutation considered targetable. In this population, 9 cases had a discordant mutational status between the primary and metastatic sites. This discordance was related to an actionable mutation in only four cases for a final concordance rate in terms of targetable alterations of 85% (22/26 patients).
Seventy-five patients underwent also a tumour molecular profile-based circulating-free DNA (cfDNA) analysis. Their characteristics are shown in Additional file 2: Table S4. Ninety-five genetic aberrations were found: 86 (90.5%) mutations, 7 (7.4%) gene copy number alterations and 2 (2.1%) fusions. Thirty-four (45.3%) patients were found to have at least one targetable mutation (median number 1; range 0-5). The most frequently altered genes are shown in Additional file 6: Figure S4. Ten patients (13.3%) were included in an EPCT, six (8%) of whom were included based on their tumour genotype profiles.
Our extensive molecular screening program allowed the identification of at least one actionable genetic alteration in 51.4% of cases and was associated with a significant clinical benefit since 27.8% of the patients in the GM group experienced a GMI > 1.3 (Additional file 2: Table S3). The low rate of technical failure and the high correlation rate between primary tumours and metastases demonstrate that FFPE archival tissue could be used effectively for molecular screening, making the need for invasive, resource-consuming and expensive tumour biopsies unnecessary. Due to tumour heterogeneity,