NENs are a group of heterogeneous tumors with distinct clinicopathological features and potentially malignant behaviors, which mostly originated from digestive system, especially the gastro-entero-pancreatic site [1–4]. As we mentioned before, H-NENs are rather uncommon and MH-NENs secondary to GEP-NENs account for their majority [6, 7, 9, 10]. With the studying goal of our research, we finally identified 150 patients with MH-NENs secondary to GEP-NENs, which had a slight male predominance with a median age of 56 yrs. (range 42-74yrs.) (Table 1). This demography was much close to the reports by some previous studies [26, 27]. In our study, 64 MH-NENs were secondary to pancreas (42.7%), 96 cases were diagnosed synchronously with the primary lesions (64%), 127 cases showed over 50% estimated liver involvement (84.7%) and 42 cases simultaneously present extrahepatic invasion at diagnosis (28%). Xiang et al. reported in their research that 44.9% of MH-NENs originated from pancreas, 65.4% of cases present synchronous liver metastasis, 82.5% of cases had over 50% estimated liver involvement and 11.1% of cases present extrahepatic invasion [27]. As a result, our estimated 3-year OS for MH-NENs secondary to GEP-NENs was 45.6%, which was a little worse than that of previous studies [26, 27], probably due to the stricter inclusive criteria that MH-NENs in the present study was just secondary to gastro-entero-pancreatic system, rather than other metastatic sites of NENs.
What’s more, with an obviously rising incidence in the past decades, a uniform pathological classification for NENs has been lacking, which substantially disabled physicians to stratify patients into prognostic groups and to guide their managements [5–7]. In 2010, WHO defined the entire group of this disease as NENs and divided them into G1 NETs, G2 NETs and “G3 NECs” based on mitotic rate and Ki-67 proliferative index [13]. The WHO 2010 grading classification for NENs was practical, whose clinical value has been validated by subsequent studies [18–21]. However, recent studies have focused on the heterogeneity of WHO 2010 “G3 NECs” group, which might consist of morphologically well-differentiated NETs with a high proliferative rate and true poorly-differentiated NECs with small-cell or large-cell features for NENs originated both from pancreas [15, 16, 25, 28] and gastrointestinal tract [29–33]. In 2017, based on some established research results on histopathologic criteria to better predict the tumor’s grade and biological behaviors, WHO redefined GEP-NENs into well-differentiated tumors of G1 NETs, G2 NETs, G3 NETs and poorly-differentiated carcinomas of G3 NECs referring mainly to both morphological differentiation and grading upon proliferation rate [22].
Till now, there isn’t a specific grading system for H-NENs (either PH-NENs or MH-NENs) due to the rarity of their epidemiology or the heterogeneity of original tumors. Considering the homogeneity of metastasis lessons and primary tumors, WHO grading classifications originally proposed for GEP-NENs have also been introduced to MH-NENs secondary to GEP-NENs [13, 14, 22]. Using the WHO 2010 grading classification, Lv et al. reported the percentage of G1, G2 and G3 among patients with H-NENs were respectively 4.94%, 25.93% and 69.13%, with a separate MST of 40.82mons, 51.87mons and 33.80mons [26]. There was significant difference on the OS when comparing G1/G2 NETs with “G3 NECs” (P = 0.011), while survival difference between G1 NETs and G2 NETs was not statistically significant (P > 0.05) [26]. Moreover, when analyzing the clinicopathological features of 112 patients with MH-NENs originating from digestive tract, Jiao et al. retrospectively enrolled 3 patients with WHO 2010 G1 NETs, 18 with G2 NETs and 91 with “G3 NECs” [34]. Interestingly, Jiao et al. identified 23 cases with G3 NETs in their “G3 NECs” category, which were morphologically well-differentiated with mean Ki-67 proliferative index of 44% (range 25%-60%) [34]. They also demonstrated that the MST of G3 NETs for patients with MH-NENs originating from digestive tract was 24.0mons, which was notably longer than 8.0mons of G3 NECs group (P < 0.01) [34].
The clinical value of the new WHO 2017 grading classification has not been rigorously validated before for patients with MH-NENs secondary to GEP-NENs. In the present study, in the light of the new WHO 2017 grading classification, we distributed all eligible patients into 4 groups for the first time, in which 10 patients with G1 NETs, 26 with G2 NETs, 33 with G3 NETs and 81 with G3 NECs were respectively identified. According to our analysis, significant differences between G1/G2/G3 NETs and G3 NECs were detected (Table 1), such as those of age at diagnosis (P = 0.041), synchronous liver lesion (P = 0.032), incidental diagnosis (P = 0.014), tumor largest diameter (P = 0.047), vascular invasion (P = 0.017) and extrahepatic metastatic disease (P = 0.029), which have resulted in the different survivals among each WHO 2017 grading group (P < 0.001; Fig. 1). Especially, survival of G3 NETs was notably better than those with G3 NECs (P = 0.012), but much worse than that of G1 and G2 NETs (P = 0.013, P = 0.037; respectively), whereas survival of G3 NECs was statistically worse than that of G1/G2/G3 NETs (P = 0.001, P < 0.001, P = 0.012; respectively). Finally, we demonstrated that synchronous liver lesion, primary tumor site, surgical margin, extrahepatic metastatic disease and WHO 2017 grading classification were effective independent predictors of OS for patients with MH-NENs secondary to GEP-NENs (Table 2). Our analyses were essentially in agreement with the results by previous studies [27, 34].
Although we firstly evaluated the prognostic relevance of this novel WHO 2017 grading classification for patients with MH-NENs secondary to GEP-NENs, we still acknowledged the limitations of our study. First of all, we confined MH-NENs to those only secondary to GEP-NENs, which will be of uncertain value for MH-NENs originated from other metastasis sites of NENs, such as lung, adrenal gland, etc. Then, the study population was relatively small, especially those with G1 or G2 NETs, which may affect the accuracy of their survival analysis. Thirdly, because the time span of our study was too long, the detailed data of each patient’s medical therapy was rather difficult to be collected, whose clinical effect would not be reflected accurately. Finally, the retrospective nature of the study prevented us from obtaining some detailed information, such as adjuvant therapy and tumor recurrence, which still needs to be further discussed.