While an accepted modality in other cancer types, such as lung cancer, maintenance therapy in gastrointestinal malignancies is not routinely used. In the setting of metastatic colorectal cancer, maintenance therapy has been recently evaluated in phase III randomized trials. For example, combination therapy incorporating capecitabine and bevacizumab, as maintenance treatment after induction therapy with capecitabine, oxaliplatin, and bevacizumab (CAPOX-B), significantly improved progression-free survival (PFS) with excellent tolerability, although the incidence of hand-foot syndrome was increased [
7]. Maintenance capecitabine monotherapy has also been evaluated after induction therapy with oxaliplatin and infusional 5-fluorouracil (FOLFOX-4) in patients with metastatic colorectal cancer with very low incidence of side effects [
8]. Little data are available for metastatic pancreatic cancer and no data for bile duct cancers. Maintenance therapy with gemcitabine after chemoradiation has been proposed for patients with locally advanced pancreatic cancer and, despite the limitation of a retrospective analysis, an increase in OS is suggested [
9]. Reure and colleagues reported a retrospective series of patients with metastatic pancreatic cancer treated with maintenance capecitabine without signs of progression after first-line leucovorin, fluorouracil, irinotecan, and oxaliplatin (FOLFIRINOX) chemotherapy with a median OS of 17 months and a median PFS of 5 months [
10]. Doherty
et al. have recently reported retrospective data suggesting a survival benefit for patients affected by biliary tract cancer receiving nine or more cycles of chemotherapy, with manageable toxicity [
11]. Here we describe the case of a women affected by metastatic ICC treated with first-line cisplatin-gemcitabine chemotherapy. In the phase III randomized trial by Valle and colleagues, the combination of cisplatin-gemcitabine was compared to gemcitabine in patients with locally advanced or metastatic biliary tract cancer [
6]. The cisplatin-gemcitabine group achieved a significantly better tumor control rate (81.4 versus 71.8%,
p = 0.049), an increased median PFS (8.0 months versus 5.0 months,
p < 0.001), and an improved median OS (11.7 months versus 8.1 months,
p < 0.001) with more incidence of bone marrow toxicity (mainly neutropenia in the combination arm). In our case, our patient achieved disease stabilization as best overall response after the first three courses of chemotherapy, with a dimensional increase in lung metastases after the sixth course, although it was not sufficient to qualify for progression. Maintenance therapy with gemcitabine was well tolerated and, surprisingly, resulted in a significant clinical delayed response, with a dimensional regression of both lung and spleen metastases. Usually, traditional chemotherapy is associated with rapid but transient rather than delayed durable response, the latter being frequently observed for novel immunotherapy drugs. It could be hypothesized that, in our case, the chemotherapy treatment had unmasked tumor antigens and the antigen release induced an immune-mediated antitumor response, without any significant change in white blood cell (WBC), neutrophil, monocyte, and lymphocytes counts. Recent data suggest that the anticancer effect of chemotherapy is also due to the interaction with innate and adaptive immune system, with various complex mechanisms [
12]. This new evidence represents the rationale to design clinical trials testing the combination of cytotoxic agents and immune checkpoint inhibitors.