TO THE EDITOR: Applying conventional diagnostic imaging paradigms, a negative 18F-FDG PET/CT study in a patient with biopsy-proven metastatic neuroendocrine tumor (NET) would be considered false-negative. With molecular imaging, however, we have emerged from using imaging merely to detect and measure lesion size to increasingly using it to characterize disease phenotype, which was formerly the domain of pathology. We therefore read with interest the recent publication by Bahri et al. (1) confirming that in patients with metastatic NET, 18F-FDG PET/CT has powerful prognostic utility superior even to conventional pathologic factors such as histologic grade or Ki-67. These data substantiate earlier data by Binderup et al. (2) and Garin et al. (3). These findings highlight the ability of PET/CT to reproducibly characterize all sites of disease in a given patient, minimizing the sampling error inherent with histopathologic sampling of a random site of disease (4).
In their prospective study, Bahri et al. (1) demonstrated a median overall survival of 15 mo for 18F-FDG–positive NET compared with 119.5 mo for 18F-FDG–negative NET. The authors also explored the additional value of somatostatin receptor imaging. In keeping with the concept that patient outcomes are related to the degree of tumor differentiation, patients with positive somatostatin receptor imaging results had a better prognosis than those without, but even so, 18F-FDG also retained its prognostic utility within this group.
The adverse prognosis associated with 18F-FDG avidity need not necessarily be the fate of such patients. We have recently published data regarding the efficacy of peptide receptor chemo-radionuclide chemoradiotherapy (PRCRT) with 177Lu-DOTATATE combined with 5-fluorouracil in a cohort of 52 patients with 18F-FDG-avid NET (5). Despite the anticipated poor prognosis of this cohort, we demonstrated an unexpectedly long progression-free survival of 48 mo, whereas median overall survival had not been reached at the time of publication. We have since updated the overall survival data of this cohort after a median follow-up of 58 mo, still with no patients lost to follow-up. Median overall survival from the commencement of PRCRT was 55 mo (Kaplan–Meier survival analysis based on log-rank testing). In response to the data presented by Bahri et al., we have further performed subanalysis in patients with a maximum standardized uptake value of at least 4.5 (n = 44) or a tumor- to normal-tissue ratio of at least 2.5 (n = 23), groups defined to have a relative risk for death of 6.2 and 23, respectively. Median survival for these subgroups in our cohort was the same as for our overall group.
These remarkable results attest to the superior efficacy of PRCRT compared with conventional therapeutic strategies, since we can assume that most patients in the study by Bahri et al. did not have access to this therapeutic modality because of lack of regulatory approval for PRRT in France, where the study was undertaken. Additionally, our results have a lead-time bias that is disadvantageous to our analysis, as survival in our study was not measured from diagnosis but rather from the time of PRCRT in a population that was previously treated with conventional therapeutic regimens, including at least one line of chemotherapy in 67%. Thus, our median survival of 55 mo is remarkable in comparison to the 15 mo defined by Bahri et al., suggesting that PRCRT prolongs survival by years in many patients with 18F-FDG–avid metastatic NET.
In addition to the encouraging results for the cohort, 4 patients have no evidence of disease after a follow-up of 30–97 mo, indicating that a small proportion of patients can be cured. Two achieved a complete response with PRCRT alone, whereas the other two were rendered disease-free after surgery; one to excise the primary site after complete regression of metastatic disease, and another in whom an R0 resection of residual primary and metastatic disease was achieved after major disease regression (6). Importantly, the resected residual disease in both patients was of significantly lower grade than that documented before treatment. Furthermore, 27% of patients in our cohort ultimately achieved a complete metabolic response on 18F-FDG PET/CT despite the presence of residual disease on somatostatin receptor PET/CT. In these patients, it appears PRCRT is able to convert the disease from an aggressive to an indolent phenotype. PRCRT is remarkably well tolerated, as we and others have previously described (5,7,8). However, there is a risk of long-term toxicity. With longer follow-up in our cohort, there have been 2 cases of myelodysplasia, although both patients remain alive after 44 and 79 mo of follow-up. This risk must be weighed against the risk of death from the underlying NET and suggests that the risk–benefit ratio is likely to be highest for patients with higher grades of NET. Although the optimal sequences for available therapies remain uncertain, we believe that the most sensible approach is to use the most efficacious and least toxic therapy upfront. For metastatic 18F-FDG–avid ENETS (European Endocrine Tumor Society) grade 2 NET, our results recommend that PRCRT be the first-line therapeutic modality of choice, and we have recently changed our multidisciplinary neuroendocrine service guidelines to reflect this recommendation.
There is further room to optimize delivery of PRCRT by refinement in patient selection and delivery of therapy (9), including the use of 90Y in patients with larger-volume disease and the use of newer chemotherapeutic combinations such as capecitabine and temozolomide for pancreatic NET (10). We are hopeful that these refinements will further improve patient outcomes.
Footnotes
Published online Mar. 26, 2015.
- © 2015 by the Society of Nuclear Medicine and Molecular Imaging, Inc.