In addition to the FDG-PET/CT findings related to the index tumor, other previously unknown malignant tumors were detected in two patients: a second primary esophageal carcinoma in one patient and a pulmonary metastasis of previously treated adenocystic carcinoma of lip in one patient. The esophageal tumor was treated with curative intent, and there has been no recurrence during the follow-up.
The aim of this study was to audit our current FDG-PET/ CT based approach to assess treatment response after definitive (C)RT for patients with newly diagnosed HNSCC. The present results of a series of 88 patients are in line with previously reported data and further confirm the reasonably good accuracy of FDG-PET/CT in this setting. The value of a negative FDG-PET/CT was particularly highlighted in the present study, with the overall NPV being 87%.
When analyzed separately for the primary site and neck, the NPV was 91% and 93%, respectively. These figures are well in line with two systematic reviews assessing the diagnostic accuracy of FDG-PET or FDG-PET/CT, in both of which the pooled NPV of FDG-PET/CT for both primary site and neck was 95%.4,5 Therefore, it seems feasible to rely on a negative FDG-PET/ CT result in the clinical follow-up of this patient population.
False positive FDG-PET/CT findings do occur, and in the aforementioned review articles, the pooled PPV figures were 75% and 59% for the primary site and 49% and 52% for the neck.4,5 We recorded better PPV figures, the overall PPV being 81%, and when analyzed separately for primary site and neck, the figures were 71% and 100%, respectively.
All the false positive findings occurred at the primary site, and thus, we did not have a single case in which FDG-PET/CT would have falsely indicated residual disease in the neck. On the other hand, the sensitivity in our study remained somewhat low being 65% in general, and 59% and 75% for the primary site and neck, respectively, which is lower than in most other studies.
Hence, a question arose whether some focal 18F-FDG accumulations representing residual tumor had been interpreted as inflammatory lesions when assessing the neck nodes. Therefore, the false negative findings in the neck nodes were re-assessed, and it was found that in one out of three such cases, different interpretation of the images would have been justifiable in a retrospective analysis.
It has been shown that optimal timing of FDG-PET/CT imaging, ie not earlier than 10–12 weeks after completion of (C)RT, is crucial to assure good accuracy.4,5 When performed earlier, accumulation of 18F-FDG may reflect treatment-induced inflammatory changes in tissues instead of residual disease. On the other hand, small traces of residual tumor may not be detected in FDG-PET/CT soon after treatment, but will become detectable a few weeks later.
This conflicts the need for prompt diagnostics and salvage treatment of patients with residual disease. It has been suggested that early surgical salvage will result in improved loco-regional control and survival rates.6 In addition, delay in performing salvage surgery may increase the risk of surgical complications. It has been proposed that there is a therapeutic window for salvage surgery at approximately six weeks after cessation of (C)RT.7
At this time point, treatment-induced acute tissue toxicity will be mostly healed, but the late toxicity will not yet be developed. The drawback of using FDG-PET/CT in monitoring treatment response at 12 weeks post-treatment is that the option for the suggested therapeutic window at 6 weeks will be lost. However, there are no means available for accurate assessment of treatment response at an earlier time point.
Despite this, it has been suggested that planned neck dissections performed a few weeks after (C)RT for patients with N2–N3 neck disease are no longer justified, because the probability of isolated neck recurrence following a complete response is very low.8,9 Even the excision of residual lymphadenopathy without FDG avidity has been questioned because of very low risk of viable tumor cells.10,11
In the current series, one patient without clinical suspicion of residual disease had persistent anatomical changes in the neck nodes in the CT portion of the imaging with no FDG avidity.
A selective neck dissection was performed and no viable tumor tissue was detected in the histopathological examination. In the case of clinically evident residual neck mass after (C)RT, it is our policy to remove the residual tumor for histopathological analysis before the scheduled FDG-PET/CT imaging.
The use of pre-treatment FDG-PET/CT is advocated by many centers.11–13 FDG-PET/CT has been shown to effectively diagnose second primary tumors and distant metastasis in HNSCC patients.12–16 Pre-treatment FDG-PET/CT also allows more targeted endoscopies and biopsies instead of panendoscopies in search for unknown or second primary tumors.
Correct staging of patients with distant metastasis is fundamental for optimal treatment planning. In addition, a baseline FDG-PET/CT is useful as a reference standard when assessing the post-treatment images for treatment response.11 Furthermore, pre-treatment FDG-PET/CT images can be used in radiotherapy target delineation.17
According to our institutional guidelines, the response assessment FDG-PET/CT should be performed at 12 weeks after cessation of the (C)RT. However, in the present retrospective analysis we found large variation in the timing of imaging, with the median interval still being approximately 13 weeks.
As discussed previously, FDG-PET/CT imaging should not be performed earlier than 10–12 weeks after completion of the treatment to assure good accuracy. On the other hand, delayed imaging at a median 16.8 weeks post (C) RT resulted in good accuracy in a study by Prestwich et al.11 The correct timing of FDG-PET/CT was not in the focus of our analysis, and we thus chose to include all patients imaged 10–18 weeks post (C)RT.