The data in the present study indicate that lower skewness (i.e., less asymmetry in the distribution) of Ktrans
may be predictive of better outcome in HNSCC patients with stage IV nodal disease. There is overwhelming evidence that tumors are heterogeneously perfused (33
). Imaging vascular heterogeneity by DCE-MRI has been shown to be useful for understanding tumor biology and predicting outcome (34
). Jackson et al. reported that tumor heterogeneity is better reflected by the distribution of the DCE-MRI parameter values than by their mean or median (34
). Studies in gliomas, breast and rectal cancers have shown that tumor heterogeneity measures, such as the upper part of the distribution (95th
percentile) or the skewness of the distribution, of DCE-MRI parameter values correlate with overall survival, tumor grade or radiation treatment outcome (35
). Tumor heterogeneity in HNSCC has been attributed in part to regions of hypoxia and necrosis within the tumor (39
). In a study of 28 HNSCC patients with stage IV disease, Brizel et al. found that tumor hypoxia (measured by pO2
) adversely affected prognosis (40
). Recently, in a study of 13 patients, our group showed that hypoxic metastatic neck lymph nodes were poorly perfused (i.e., had significantly lower kep
values) compared with nonhypoxic nodes (30
). Additionally, hypoxic nodes had a more asymmetric distribution of kep
values than did non-hypoxic nodes (30
). These thirteen patients were also included in the present study.
In advanced HNSCC, tumor stage and neck node involvement are widely recognized as negative prognostic factors (8
). Our results have shown that in Cox regression analysis, only T-stage and either age or sex had a jointly significant relationship to PFS or OS. The present study examined imaging measurements from neck nodal metastases only. The primary endpoints of the analysis in the present study were PFS (including assessment of all tumor sites) and OS, rather than just disease progression at the primary site. Our results show that information about tumor vascularity from a priori DCE-MRI may help in stratifying HNSCC patients with stage IV nodal disease into for risk-adjusted treatment selection.
DCE-MRI parameters have shown promise as early markers of response in HN cancers (20
). Cao, et al. performed quantification of blood volume (BV) and blood flow (BF) from DCE-MRI data before therapy and 2 weeks after initiation of chemoradiation in 14 HNSCC patients; they found that an increase in available primary tumor BV during RT was associated with local-regional control (26
). The median follow-up for the ten surviving patients in their study was 9.7 months (range, 5.3–27 months) (26
). In a cohort of 33 HNSCC patients who were treated with chemoradiation, Kim, et al. found that the average pretreatment Ktrans
value of the complete response group was significantly higher (P = 0.001) than that of the partial response group at 6-month follow-up (20
). Our study has a larger number of patients (n=74 patients) and longer clinical follow up (minimum one year).
The use of functional imaging (with CT, PET or MRI) is gaining acceptance in the management of patients with HN cancers (43
). DCE-MRI measures changes in signal intensity while perfusion CT measures changes in tissue attenuation during a dynamic contrast infusion (43
). In a study of 105 HNSCC patients who underwent perfusion CT followed by radiotherapy, Hermans, et al. showed that patients with lower median perfusion values had a significantly higher local failure rate (p<0.05) (44
). The results confirmed the hypothesis that less-perfused tumors respond poorly to radiotherapy (44
). Consistent with this hypothesis, we found that the skewness of the perfusion parameter Ktrans
was higher in stage IV patients who had heterogeneous, poorly perfused tumors and died of disease thereafter.
A few studies have shown that 18
F-FDG) uptake intensity in HNSCC has some prognostic value (45
). These studies have used the maximum standardized uptake value (SUVmax) as a predictor of outcome (45
). Perfusion CT and PET, unlike DCE-MRI, use ionizing radiation, which has known risks. Efforts have been initiated by the National Cancer Institute Cancer Imaging Program to develop DCE-MRI as a mainstay of diagnostic imaging that can provide quantitative data for use in multicenter clinical trials (48
). DCE-MRI is rapidly gaining acceptance as a tool for early assessment of therapeutic response in clinical trials (18
). In a Phase II study of HNSCC patients treated with sunitinib, a significant decrease in Ktrans
was seen in three of the four patients who received DCE-MRI monitoring (21
Our study has a few limitations. First, the study design was retrospective, and the patients were not consecutive. Second, patients received different treatments; however, this reflects daily practice at the Center. Thirdly, we had HPV status on 48 (65%) patients; it would have been better if we had the data for the whole population, particularly since the oropharynx was the predominant primary site. Finally, our study included no direct pathological validation of heterogeneity (necrosis). Future prospective validation studies are needed to confirm our findings in a larger patient population and known predictors that were not significant in the present study should be pursued.