Several recent innovations have been introduced in efforts to improve the clinical outcome in the treatment of HNC. New radiotherapy technologies such as IMRT and IGRT have improved radiotherapy delivery and new surgical techniques including sentinel lymph node biopsy also have potential for improving HNC therapy. In addition, increased implementation of diagnostic exams such as PET and PET/CT scans have the potential to improve treatment outcomes by providing improved lymph node staging, prognostic, and perhaps predictive factors.
We found excellent positive and negative predictive value for PET scans in determining lymph node status. This finding is in line with several other reports [5
]. Table . These findings have potential clinical implication in that it may be possible to use staging PET scan results in stratifying patients who should go on to receive neck dissections. Ng et al. [11
] found false negative PET scan results depended on the T stage of the tumor with false negative PET scans in 6.7, 10.8, 13.3, and 25% for T1-4 tumors respectively. Our results also demonstrate the low false negative rate in staging PET scans, although we did not find a relation between T stage and false negative rates, likely secondary to the low number of false negative PET scans. A recent meta-anaylsis [28
] examining the results of PET scan in cervical lymph node staging found lower sensitivity (79%) and specificity (86%) results than our study. Some explanations for this include the fact that most of the studies in the meta-analysis were older and used PET scan alone rather than combined PET/CT.
Results for Staging with PET
Several authors have examined if PET/CT is superior to PET alone in nodal staging [29
]. Bransetter et al. [30
] reported that PET/CT was superior to PET alone in HNC lesions, and Schoder et al. [32
] arrived at the same conclusion. The overall accuracy in determining lymph node status was not statistically significantly better for PET/CT in this study, but there was a trend towards significance. Although PET/CT had several false-positive results there were no false negatives. The added accuracy of CT fusion with the PET scan, especially in terms of reducing the number of false negative scans, is important in the use of PET/CT to limit surgical neck dissection.
This report and others can help determine the best use for PET scans in the nodal staging of HNC. Our report lends strength to the previously published studies regarding the excellent predictive value of PET scans in determining lymph node status. Having adequate knowledge of lymph node status without the aid of surgical dissection has several potential areas of clinical benefit. Neck dissections add significant morbidity in HNC patients [33
] and PET scans may be able to determine which patients do not require a full dissection. If further evidence, perhaps in the form of a prospective trial, can confirm the accuracy of PET scans in lymph node staging, it would be possible to use imaging in lieu of surgery for determining lymph node status.
The role of maximum SUV as a prognostic factor is controversial. Several reports have found a particular SUV value associated with a worse outcome [18
]. However, other authors did not report this association [26
]. One factor in the mixed findings of maximum SUV and outcome may be related to the relatively small number of patients in many of these studies. Our data found that SUV of the tumor mass > 8 was associated with worse overall survival. Interestingly, we did not find an association between tumor mass SUV and local control, likely secondary to several late relapses. Fewer reports have focused on the prognostic information of nodal SUV. Schwartz et al. [19
] did not find any association between nodal SUV and outcome in 36 patients, Brun et al [20
] also failed to find any association and also had 36 patients. Liao et al. [27
] examined only oral cavity patients with positive lymph nodes and found that a lymph node SUV of 5.4 predicted for worse outcomes. We found an association between maximum SUV of the lymph node and outcome with nodal SUV > 10 more likely to have distant failure in 79 patients with nodal SUV. One explanation of the discrepancy between our data and that of Schwartz et al. and Brun et al. could be the larger number of patients which would give our study power to detect a true finding.
Although prognostic information, such as tumor mass SUV, is useful, it is not used for management changes. However, the association between maximum lymph node SUV and distant failure has possible clinical implications. One of the current trends (although still unproven) in HNC is neoadjuvant chemotherapy with TPF [34
]. If further reports confirm the finding that maximum lymph node SUV > 10 has a higher likelihood of distant failure but not a higher likelihood of local failure, this would be a possible stratification factor for which patients should receive neoadjuvant TPF rather than concurrent therapy.
We also found that maximum SUV was predictive of ECE. ECE is an established risk factor and is one of the indications (the other being positive margins) in determining which patients post-operatively benefit from combined chemoradiotherapy rather than radiotherapy alone [36
]. For patients receiving definitive radiotherapy, it may make be beneficial to treat higher SUV lymph nodes (which are more likely to harbor ECE) to a higher dose.
The results of this study need to be interpreted with caution; it is retrospective in nature and includes a variety of head and neck cancer subtypes. In this regard, our study has the same flaws as the majority of other reports examining the utility of FDG-PET in the staging of head and neck cancer. However, this study does have some improved clinical utility secondary to its size (212 PET scans analyzed) and the multi-institution nature of this report. We acknowledge that most of the information gathered are hypotheses generating and that further studies, perhaps in a prospective randomized fashion, will be required to truly determine the utility for PET scans in HNC. Until such studies are completed, large retrospective such as this are needed to provide the best information at this time possible.