Tumori Journal Abstract Book - October 2020 - 12

Tumori Journal 106(2S)

FFPE tumor blocks and/or liquid biopsies in a multicenter
compassionate program using Foundation 1 Medicine
(FM1) platform or Guardant 360 test. The aim of our analysis was to investigate the presence of potential druggable
alterations and their correlation ù with baseline clinical
factors.
Results: Overall, thirty-four aNSCLC patients were gathered. Median age was 64 years with 53% of male versus
47% of female. Never smokers were 11 patients (32%)
versus 23 patients former/current smokers (68%). 95%
had an ECOG Performance Status of 0-1. 74% were metastatic at diagnosis with 88% of adenocarcinoma, 6% of
squamous carcinoma and 6% not otherwise specified
(NOS) cancers. Complete molecular results were available in 30 patients (88%), as test failed in 4 patients (12%)
due to inadequate materials (tissue for 3 patients and liquid biopsy for 1 patient). Among 30 patients with available results, a targetable alteration was detected in 5
patients (17%). Among these, 1 was EML4/ALK fusion
(local ICH test was negative), 1 KIF5B-RET fusion, 1
BRAFV600E mutation, 1 MET exon 14 skipping alteration, 1 EGFR exon 18 E709_T710>D mutation. For all
these patients a target treatment was started. No targetable
alterations were found in smoker patients, as all targetable
alterations were found in non-smokers patients in which
molecular test was available (10/11). Smoking status
demonstrated a significant differential effect (Chi-square
test, p=0.009) in predicting the presence of druggable
alterations. Among no targetable alterations of interest,
KRAS was mutated in 30% (9/30) of the patients (4 of
them had concomitant STK11 mutations), both ERBB2
alterations (A730T and P780-Y781insGSP) and FGFR
family alterations in 7% (2/30). Incidence of targetable
alterations among KRAS wild-type patients was not statistically different from the overall population (24% versus 17%, chi-square p=0.731).
Conclusions: Our analysis suggests that NGS analyses
may provide useful results in aNSCLC patients. Since
clinically significant results were obtained only in nonsmoker patients, we suggest considering smoking status as
a possible driver for NGS analyses in aNSCLC.

Background: Currently, predictive factors for recurrence
risk of resected, early stage non-small cell lung cancer
(NSCLC) are limited to post-surgical stage, and novel predictors might improve patient selection for adjuvant chemotherapy. Recently, circulating tumor-derived exosome (Exo)
miRNAs have emerged as highly specific non-invasive biomarkers for early NSCLC diagnosis. In this study, we aimed
at identifying an Exo-miRNA signature to select resected
NSCLC patients at higher risk of disease recurrence.
Material and methods: The expression profiles of 2,549
miRNAs were screened by microarray in the plasma ExomiRNome isolated from 67 patients, who underwent resection for NSCLC and had a follow-up of at least 5 years. The
plasma for Exo-miRNome analysis was collected before
surgery. To build a predictive score, we applied penalized
Cox regression analysis, with the LASSO method, to a subset of miRNAs with mean logIntensity higher than 5, age,
sex and stage. Finally, we plotted the Kaplan-Meier curves
of patients divided according to the median predictive score
and tested their difference using the Log Rank test.
Results: The median age of patients was 68 years, of
whom 76% were men. The disease stages were classified
as follows: stage I (42%), stage II (46%), stage III (12%).
Disease recurrence (interval: 4-75 months) was observed
in 36/67 (54%) patients. A patient "risk score" was built by
combining the stage with 5 (miR-4481, miR-4436b-3p,
let-7b-3p, let-7e-5p, miR-3620-3p) and 3 (miR-4716-5p,
miR-1249-3p, miR-6766-3p) Exo-miRNAs, which were
either positively or negatively associated with disease
recurrence, respectively. Differences of both recurrencefree survival and overall survival of patients' groups were
statistically significant (p< 0.0001 for both) according to
the predictive risk score. In particular, high risk patients
had median recurrence-free survival 4 times lower than
low risk patients.
Conclusions: The risk score we found, which include 8
Exo-miRNAs and disease stage, can predict early stage
NSCLC patients, who are more likely to relapse after initial surgery. Such patients may eventually be eligible for
personalized protocols involving post-operative chemo/
radiotherapy.

A15

A16

Prediction of recurrence in
resected non-small cell lung
cancer by means of an exosomal
miRNA-based risk score

RELAY, erlotinib plus ramucirumab
or placebo in untreated EGFRmutated metastatic NSCLC:
outcomes by EGFR mutation type

Rossi G.1, Coco S.1, Longo L.1, Chiorino G.2, Ostano P.2, Dal Bello
M.G.1, Grassi M.1, Venturi C.1, Mastracci L.3, Tagliamento M.1,
Dellepiane C.1, Zullo L.1, Beshiri K.1, Alama A.1, Bennicelli E.1, Bottini
A.1, Pronzato P.1, Genova C.1

Nakagawa K.1, Nadal E.2, Garon E.3, Nishio M.4, Seto T.5,
Yamamoto N.6, Park K.7, Shih J.8, Frimodt-Moller B.9, Zimmerman
A.10, Visseren-Grul C.11, Reck M.12, Pasello G.P.13

IRCCS Ospedale Policlinico San Martino, Genova; 2Fondazione Edo ed Elvo
Tempia, Biella; 3Università degli Studi di Genova, Genova

1
Department of Medical Oncology, Kindai University Faculty of Medicine,
Osaka; 2Department of Medical Oncology, Catalan Institute of Oncology,
L'Hospitalet de Llobregat, Barcelona; 3David Geffen School of Medicine at

Tumori Journal Abstract Book - October 2020

Table of Contents for the Digital Edition of Tumori Journal Abstract Book - October 2020