Background
The prognosis of patients with advanced esophageal cancer continues to be poor, despite advances in management. Neoadjuvant chemotherapy or chemoradiation therapy before esophagectomy is a standard-of-care and commonly applied in clinical practice for locally advanced and operable esophageal cancer [
1‐
3]. Once neoadjuvant therapy is completed, assessment of response is necessary [
3]. When a persistent local lesion is indicated, esophagectomy is strongly recommended because the presence of residual tumor after neoadjuvant therapy in the resected specimen leads to shorter overall survival [
3‐
5]. On the contrary, if there is no evidence of residual viable lesion, surveillance can be a possible option [
3]. Generally, FDG PET/CT and/or contrast-enhanced chest CT is used for the evaluation of treatment response of neoadjuvant therapy in esophageal cancer [
3]. It is sometimes difficult to distinguish a viable residual tumor form reactive changes with a chest CT. In contrast, FDG PET/CT provides a more accurate diagnosis compared to that with chest CT due to its evaluation of metabolic activity. However, the value of FDG-PET/CT for evaluating response to neoadjuvant therapy in esophageal cancer is still controversial [
6‐
8], so it is basically not recommended for the selection of patients for esophagectomy following neoadjuvant therapy [
3].
Recently, Toyohara et al. developed 4′-[methyl-11C] thiothymidine (4DST) as a new DNA synthesis imaging agent [
9,
10]. Although 3′-fluoro-3′-deoxythymidine (FLT) has been established as a cell proliferation PET tracer, 4DST has advantages for proliferation measurement [
11]. 4DST incorporates into DNA directly, whereas FLT does not incorporate into DNA and reflects salvage pathway of DNA synthesis [
12]. We have previously reported on the great potential of 4DST PET/CT for proliferation imaging in malignancies such as lung cancer and renal cell carcinoma [
13‐
15]. In addition, Hoshikawa et al. have reported that 4DST PET shows a higher prognostic value in patients with head and neck carcinoma compared to FDG PET [
16].
These results suggest that 4DST PET can potentially predict a response to neoadjuvant therapy in esophageal cancer. The aim of this study was to evaluate the diagnostic value of 4DST for predicting response to neoadjuvant therapy in patients with esophageal cancer as compared to that with FDG.
Discussion
The purpose of this study was to investigate whether 4DST-PET is useful for predicting the treatment response in patients with esophageal cancer, compared to the usefulness of FDG-PET. 4DST ΔSUVmax provided high AUC values to distinguish treatment responders among all PET parameters. In addition, 4DST postSUVmax was also helpful for predicting response.
While FDG
postSUVmax was not statistically different between responders and non-responders, 4DST
postSUVmax were statistically higher for responders and showed great diagnostic performance in the discrimination of responders (AUC = 0.88). It has been reported that post-therapeutic FDG-PET is useful for evaluation of tumor response in esophageal cancer [
23‐
25], which is concordant with our results. A possible explanation why 4DST
postSUVmax showed greater diagnostic value than that of FDG is that FDG accumulation can be affected by inflammation due to chemotherapy or chemoradiation therapy [
26,
27], whereas the influence of these therapies for 4DST may not be as significant. Indeed, high tumor selectivity of 4DST, which enables discrimination between tumor and inflammation, has been demonstrated in the rodent model [
28]. The other possible reason is that 4DST simply measures tumor proliferation more accurately than FDG does. 4DST uptake corresponded well to Ki-67 [
29], and in lung cancer, Minamimoto et al. have reported 4DST shows a better correlation with Ki-67 than FDG does [
13]. Therefore, 4DST may have the potential to reflect tumor viability more precisely than FDG, not only in lung cancer but also in esophageal cancer. The higher accumulation of 4DST is likely to indicate more residual viable cancer cells.
Both FDG ΔSUVmax and 4DST ΔSUVmax were useful for evaluating response, and PERCIST 1.0 also provided good diagnostic performance, whereas 4DST ΔSUVmax showed higher accuracy than PRECIST 1.0. It has been reported that percent changes of FDG uptake (including PERCIST 1.0) between pre- and post-treatment are effective for the evaluation of response in esophageal cancer [
21,
23,
30]. In PERCIST 1.0, the cutoff value for the response is set at more than or equal to 30%. However, several studies suggested optimal cutoff values for response evaluation in esophageal cancer as high as 50–60% [
21,
23,
31]. This percentage was similar to optimal cutoff values of FDG ΔSUVmax in our study (60.3%), which showed higher diagnostic performance than when using 30% as a cutoff value. Indeed, it is frequently difficult to determine the optimal cutoff value in esophageal cancer, because non-responders also tend to show reduction of FDG accumulation [
22,
23,
30]. In fact, the reduction rate of FDG ΔSUVmax in non-responder was not small (median FDG ΔSUVmax = − 36.3%) in this study as well. Conversely, in 4DST PET, non-responders showed almost no change from baseline PET, whereas responders showed a great reduction of uptake (median 4DST ΔSUVmax: non-responder = − 2.9%, responder = − 56.7%). These results indicate that persistent 4DST uptake into primary lesion after treatment is highly suggestive of inadequate response. As such, the percent changes of 4DST accumulation are likely to be an easy-to-use marker in clinical practice.
FDG
preSUVmax was higher for responders than for non-responders, which was statistically significant (
p = 0.018). Although various studies have failed to demonstrate the prognostic value of the baseline SUVmax in FDG-PET [
26,
32‐
34], some reported higher FDG SUVmax in initial PET correlated with a higher rate of complete histological response [
35], and even with better disease-free survival [
36]. Some suggested this is because higher FDG SUVmax is related to higher proliferation cells that are rapidly proliferating may better respond to chemotherapy or chemoradiation therapy [
37]. However, if this theory is true, 4DST
preSUVmax of responders should have been statistically higher than that of non-responders. This point should be discussed in future studies.
Further, based on ΔSUVmax, the ability for the differentiation of responders from non-responders were almost identical as depicted in ROC (Fig.
2); however, a few cases showed discrepancy between FDG and 4DST ΔSUVmax (
n = 2). As discussed earlier, a possible explanation for this discrepancy could be that FDG accumulation can be affected by inflammation due to chemotherapy or radiation therapy, whereas the influence of these therapies for 4DST may not be significant. This issue needs to be discussed in further studies with large sample size.
This is the first study to report the utility of 4DST for the evaluation of therapeutic response in malignancy. Gerbaudo et al. has reported the efficacy of FLT, of which biodistribution is like 4DST [
17], for the evaluation of treatment response in esophageal cancer [
38]. Although the number of patients was limited in their study, the authors first showed that in patients with esophageal adenocarcinomas, FLT-PET demonstrated early therapy-induced decrease in tumor proliferation in response to treatment. The difference in the reduction in tumor FLT uptake during treatment between responders and non-responders was statistically significant, but for FDG, it was not. This was probably due to the fact that FLT uptake was not affected as much as FDG was by radiation induced inflammation. On the other hand, the difference in the reduction in tumor FLT uptake at after completion of therapy between responders and non-responders was no longer significant. Gerbaudo and colleagues explained that at the end of chemoradiation treatment, FLT uptake could have been affected by the continuous effect of radiation, which diminished the proliferative capacity of remaining viable cells. In contrast,
postSUVmax and ΔSUVmax in 4DST were useful to distinguish responders from non-responders at the end of therapy in our study. The advantage of 4DST over FLT in proliferation measurement is presented in vivo analysis [
10]. Namely, FLT is not incorporated into DNA and reflects salvage pathway of DNA synthesis [
12], whereas 4DST incorporates into DNA directly [
9]. This feature may be one of the possible reasons to explain the discrepancy of usefulness between 4DST and FLT-PET in the assessment of treatment response. The other reason might be that more than half of the patients in our study were not treated with radiation. Considering Gerbaudo et al. findings described above about the use of interim FLT-PET for early therapeutic monitoring in esophageal cancer, it is probable that interim 4DST-PET will also have great diagnostic potential in this setting and should be examined in future studies.
There are some limitations in this study. Firstly, it was performed in a single center with a relatively small number of patients. Larger prospective multicenter trials are necessary in the future. Secondly, the mixed population of patients in terms of neoadjuvant therapy (chemotherapy vs. chemoradiation therapy) is another limitation, but this factor was not statistically significant between responders and non-responders. Thirdly, three PET scanners were used in this study, which potentially influenced the values of the PET parameters. However, cross-calibration between the three scanners was performed, and the same scanner was used in each patient not only for pre- and post-neoadjuvant therapy scan but also for 4DST and FDG PET scan. Thus, the influence of the difference of scanners is considered minimal. Finally, a cyclotron is necessary for the production of 4DST, which may be a drawback for this C11-labeled PET tracer.