Letter to the editor

Previous precision medicine studies have investigated conventional molecular techniques and/or limited sets of gene alterations [1–3]. 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.

In 28 cases (5%), molecular analysis failed mainly because of insufficient tissue quantity or quality. The median time from first referral to reporting was 9 weeks (range 1–36 weeks). The 20 genes found 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 (Additional file 4: Figure S2). One thousand and six hundred fifty-nine alterations were found: 883 mutations (53.2%), 755 (45.5%) gene copy number alterations and 21 (1.3%) fusions. The median number of alterations per patient was 2 (range 0–18).

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.

One hundred fifty-nine (28%) patients were randomized in an early phase clinical trial (EPCT) after the screening results. The main reasons for non-inclusion were non-progressive disease on current treatment regimen (31.5%), general status deterioration (25%), death (16.5%), clinical trial not available (10.5%), screening failure (6.5%), loss to follow-up (7%) and patient refusal (3%). The drug used in the EPCT was genotype-matched (GM) in 86 (15.1%) patients and non-matched (NM) in 73 (12.9%) patients (Fig. 1). The drugs received are summarized in Additional file 2: Table S2.

Fig. 1
figure 1

Study flow diagram

Full size image

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 progression-free survival (PFS) was 3 months. The median overall survival was 8.5 months (range 5.5–11.5 months).

Fifty-nine patients were evaluable for the growth modulatin index (GMI) calculation. The median GMI [4] was 0.63 (0.01–5.81). Twenty patients (27.8%) had a GMI ≥ 1.3 (Additional file 2: Table S3). A GMI ≥ 1.3 was associated with a trend towards improved median overall survival: 11.7 months (range 0.3–23.1) versus 7.6 months (range 4.8–10.5 months) for GMI < 1.3, which was not statistically significant (p = 0.28).

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, biopsies often suffer from sample bias and archival tissue is not always available. In this regard, we report here for the first time the value of an NGS assay targeting 20 cancer genes to detect actionable mutations and rearrangements in cfDNA in the context of a precision medicine study.

Overall, this study demonstrates the feasibility and potentially positive clinical impact of using comprehensive molecular profiling to improve the outcomes of cancer patients.