Page 160 - Quantitative Imaging of Small Tumours with Positron Emission Tomography

For nodal staging using SUV, non-significant trends of increased accuracy for breast, head-and-neck squamous cell, and thyroid cancer (from 80%, 66% and 95% to 84%, 71% and 100%, respectively) (26-28), and decreased accuracy for nasopharyngeal and prostate cancer (from 84% and 85% to 73% and 80%, respectively) were observed (29,30). The study investigating accuracy of nodal staging of nasopharyngeal cancer did observe a large increase in accuracy, from 14% to 71%, when stratifying for lesion size (6-7mm diameter) (29). With visual image interpretation, PSF reconstruction tended to increase accuracy of nodal staging in NSCLC, breast, and colorectal cancer (not statistically significant) compared to non-PSF reconstruction (from 76%, 76%, and 89% to 84%, 80%, and 92%, respectively) (31-33). Another study found no significant difference in lung cancer (several types) overall staging accuracy between non- PSF and PSF reconstruction (34). Prognosis Impact of PVC on prognostication (Table 3, n=6) was investigated for NSCLC (n=2), esophageal (n=2), and head-and-neck cancer (n=2). Applied PVC methods were the recovery coefficient-method (n=4), iterative deconvolution (n=1), and mask-based PVC (n=1). Only prognostic studies providing survival data were 7 included. One study did not specify lesion sizes. None of the studies stratified results on PVC for lesion size in secondary analysis. PVC did not alter the association of SUVmax with disease-free survival of NSCLC (various histological types) patients in multivariate analysis (35,36). Similarly, in NSCLC patients (various histologic types) PVC did not alter the ROC area-under-the-curve of primary tumour SUVmax to differentiate between groups of patients in terms of disease-free and overall survival (36). Primary tumour SUVs, regardless of PVC, were insufficient as prognostic markers in esophageal (adeno- and squamous cell) cancer in univariate and ROC analysis (37,38). In head-and-neck cancer patients, partial-volume-corrected SUV was significantly different between patient groups stratified according to disease-free survival, whereas uncorrected SUV was not (39). In univariate analysis, PVC did not affect predictive value of head-and-neck cancer primary tumour SUV on local recurrence-free survival, distant metastasis-free survival, and disease- free survival, but did allow for prediction of distant metastasis-free survival in a subgroup of patients with PET-positive lymph nodes (40). Systematic review and meta-analysis 159