Early changes in plasma DNA levels of mutant KRAS as a sensitive marker of response to chemotherapy in pancreatic cancer

www.nature.com/scientificreports OPEN Received: 4 April 2017 Accepted: 6 July 2017 Published: xx xx xxxx Early changes in plasma DNA levels of mutant KRAS as a sensitive marker of response to chemotherapy in pancreatic cancer Marzia Del Re1, Caterina Vivaldi2, Eleonora Rofi1, Enrico Vasile2, Mario Miccoli3, Chiara Caparello2, Paolo Davide d’Arienzo2,4, Lorenzo Fornaro2, Alfredo Falcone2 & Romano Danesi1 Pancreatic cancer (PDAC) is still lacking of reliable markers to monitor tumor response. CA 19-9 is the only biomarker approved, despite it has several limitations in sensitivity and specificity. Since mutations of KRAS occur in more than 90% of tumors, its detection in circulating free tumor DNA (cftDNA) could represent a biomarker to monitor chemotherapy response. Twenty-seven advanced PDAC patients given first-line 5-fluorouracil, irinotecan and oxaliplatin or gemcitabine and nab-paclitaxel were enrolled. Three ml of plasma were collected: 1) before starting chemotherapy (baseline); 2) at day 15 of treatment; and 3) at each clinical follow-up. cftDNA was extracted and analysed for KRAS mutations (mutKRAS) by digital droplet PCR. Nineteen patients displayed a mutKRAS in baseline plasma samples. There was a statistically significant difference in progression-free survival (PFS) and overall survival (OS) in patients with increase vs. stability/reduction of cftDNA in the sample collected at day 15 (median PFS 2.5 vs 7.5 months, p = 0.03; median OS 6.5 vs 11.5 months, p = 0.009). The results of this study demonstrate that cftDNA mutKRAS changes are associated with tumor response to chemotherapy and support the evidence that mutKRAS in plasma may be used as a new marker for monitoring treatment outcome and disease progression in PDAC. Pancreatic adenocarcinoma (PDAC) represents the fourth leading cause of cancer-related deaths in Western countries and only 10–20% of patients are diagnosed with a resectable disease1. Chemotherapy represents the milestone of treatment of advanced disease; indeed, among the combinations developed in recent years, 5-fluorouracil, irinotecan and oxaliplatin (FOLFIRINOX), and the association of gemcitabine and nab-paclitaxel (GEMnPAC) have demonstrated improved survival and represent now the standard of care in first-line treatment2, 3. However, since pancreatic tumor tissue is surrounded by a dense fibrotic stroma, the evaluation of tumor response to therapy is particularly challenging and imaging techniques do not always provide an accurate estimate, despite the significant costs4, 5. For this reason, the monitoring of these tumors by blood markers is a valuable alternative and is based on the evaluation of the Carbohydrate Antigen 19-9 (CA 19-9) levels, the standard serum marker for pancreatic cancer. Although CA 19-9 has been approved to monitor tumor response, it has several limitations including: (1) low sensitivity and specificity (estimated to be around 79% and 82%, respectively); (2) occurrence of unspecific changes in serum samples; (3) poor accuracy in the identification of small tumors6, 7. These issues limit the utility of CA 19-9 as a biomarker and highlight the need for new biomarkers to complement the imaging in order to obtain a more effective monitoring of these patients and improve the clinical outcome. Alternative approaches for PDAC monitoring are represented by circulating tumor cells (CTCs), which may predict poor disease outcome8, though their use is limited by false negative tests and high costs. 1 Unit of Clinical Pharmacology and Pharmacogenetics, Department of Clinical and Experimental Medicine, University of Pisa, Pisa, Italy. 2Medical Oncology Unit, Department of Translational Research and New Technologies in Medicine, University of Pisa, Pisa, Italy. 3Department of Clinical and Experimental Medicine, University of Pisa, Pisa, Italy. 4Sant’Anna School of Advanced Studies, Department of Medical Sciences, Pisa, Italy. Marzia Del Re and Caterina Vivaldi contributed equally to this work. Correspondence and requests for materials should be addressed to M.R. (email: ) SCIENTIfIC REPortS | 7: 7931 | DOI:10.1038/s41598-017-08297-z 1 www.nature.com/scientificreports/ N (%) Total 27 (100%) Gender Male 14 (52%) Female 13 (48%) Age Median (range) 68 (49–77) ECOG Performance Status 0 23 (85%) 1 4 (15%) Chemotherapy regimen FOLFIRINOX 13 (52%) GEMnPAC 14 (48%) Stage III 4 (15%) IV 23 (85%) Primary tumor location Head 13 (48%) Body-tail 14 (52%) Baseline CA 19-9 Normal (0–37 U/ml) 7 (26%) Abnormal (<59 ULN) 11 (41%) Abnormal (≥59 ULN) 9 (33%) Baseline KRAS cftDNA status Wild type 8 (29.6%) KRAS 19 (70.4%) mut KRAS mutations p.G12D 14 (73%) p.G12R 2 (11%) p.G12V 2 (11%) p.G13D 1 (5%) Median baseline mutKRAS (copies/ml, range) 2100 (80–64800) Table 1. Characteristics of patients. The majority of pancreatic cancers (from 75 to 95%) harbor a KRAS mutation at codon 129 and a significant difference in survival was observed in patients with detectable or undetectable KRAS mutations (mutKRAS) in circulating free tumor DNA (cftDNA)10. Moreover, the presence of mutKRAS in plasma samples after surgical resection of PDAC has been associated with poor survival11, 12. In addition, looking at the specific type of mutation, the presence of KRAS p.G12V seems to be associated with a decrease in survival of pancreatic cancer patients13. Tissue availability is recognized as a crucial issue, as there are several limitations to obtain repeated biopsies (i.e. invasiveness) and to capture tumor heterogeneity due to the practical constraints to collect multiple samples. In contrast, the detection of somatic mutations in cftDNA released in plasma from apoptotic and necrotic tumor cells both from primary tumor and metastatic lesions is feasible and affordable in terms of costs14. Thus, cftDNA may reflect the tumor burden in cancer patients and is a valuable option for a better monitoring of response; it may complement imaging in order to obtain a more effective management of patients, identify early resistance and improve clinical outcome15. Therefore, the aim of the present study was to monitor the allelic burden of mutKRAS in patients with advanced PDAC undergoing first-line combination chemotherapy. In particular, the correlation between early variations of cftDNA mutKRAS after 15 days of treatment with response to therapy was explored, in order to provide an effective tool to predict the effectiveness of therapy early after its beginning. Results Study population. A total of 27 patients with locally advanced (n = 4; 15%) and metastatic (n = 23; 85%) PDAC were included in this biomarker study (Table 1). Median age was 68 years (range 49–77). At a median follow up of 7.6 months, 13 patients (54%) underwent disease progression and 6 (22%) died. Median progression-free survival (PFS) and overall survival (OS) were 7.4 and 11.5 months, respectively. Response rate was 38% and disease control rate (DCR) d (...truncated)