Pre- and per-treatment 18F-FDG PET/CT parameters to predict recurrence and survival in cervical cancer.

PURPOSE: To describe the evolution and to assess the predictive value of metabolic parameters with different SUV threshold segmentations calculated from two 18F-FDG-PET/CT, one prior to and the other one during concomitant chemoradiation therapy (CCRT), for locally-advanced cervical cancer (LACC). MATERIAL AND METHODS: 53 patients treated for LACC by CCRT underwent FDG-PET/CT before treatment (PET1) and another one at 40Gy (PET2). The PET analyzed parameters were: maximum standardized uptake value (SUVmax), metabolic tumor volume (MTV), and total lesion glycolysis (TLG). MTVs were automatically segmented using various percentage SUVmax thresholds (30-70%) and fixed SUV thresholds (all voxels with SUV >1-20). The parameters' predictive capabilities for disease-free (DFS) and overall survival (OS) were assessed using the Harrell's C-index (c) and Cox regression model. RESULTS: Depending on the SUVmax threshold, the relative decreases in MTV and TLG from PET1 to PET2 were, on average, 65% (range: 63-70%) and 85% (range: 83-86%), respectively. The strongest predictive threshold segmentations were 55% SUVmax in PET1 and 32% in PET2. Significant predictors of DFS in multivariate analysis (c=0.82) were MTV1 (55% SUVmax) in PET1 and TLG2 (32% SUVmax) in PET2. MTV1 (55%) was the most significant OS predictor. CONCLUSIONS: MTV and TLG calculated with a threshold of 55% SUVmax and 32% SUVmax from pre- and per-treatment PET scans respectively, can be used to predict patient outcome after CCRT for LACC.

Detection of bladder metabolic artifacts in (18)F-FDG PET imaging.

Positron emission tomography using (18)F-fluorodeoxyglucose ((18)F-FDG-PET) is a widely used imaging modality in oncology. It enables significant functional information to be included in analyses of anatomical data provided by other image modalities. Although PET offers high sensitivity in detecting suspected malignant metabolism, (18)F-FDG uptake is not tumor-specific and can also be fixed in surrounding healthy tissue, which may consequently be mistaken as cancerous. PET analyses may be particularly hampered in pelvic-located cancers by the bladder׳s physiological uptake potentially obliterating the tumor uptake. In this paper, we propose a novel method for detecting (18)F-FDG bladder artifacts based on a multi-feature double-step classification approach. Using two manually defined seeds (tumor and bladder), the method consists of a semi-automated double-step clustering strategy that simultaneously takes into consideration standard uptake values (SUV) on PET, Hounsfield values on computed tomography (CT), and the distance to the seeds. This method was performed on 52 PET/CT images from patients treated for locally advanced cervical cancer. Manual delineations of the bladder on CT images were used in order to evaluate bladder uptake detection capability. Tumor preservation was evaluated using a manual segmentation of the tumor, with a threshold of 42% of the maximal uptake within the tumor. Robustness was assessed by randomly selecting different initial seeds. The classification averages were 0.94±0.09 for sensitivity, 0.98±0.01 specificity, and 0.98±0.01 accuracy. These results suggest that this method is able to detect most (18)F-FDG bladder metabolism artifacts while preserving tumor uptake, and could thus be used as a pre-processing step for further non-parasitized PET analyses.

Random forests to predict tumor recurrence following cervical cancer therapy using pre- and per-treatment 18F-FDG PET parameters.

The ability to predict tumor recurrence after chemoradiotherapy of locally advanced cervical cancer is a crucial clinical issue to intensify the treatment of the most high-risk patients. The objective of this study was to investigate tumor metabolism characteristics extracted from pre- and per-treatment 18F-FDG PET images to predict 3-year overall recurrence (OR). A total of 53 locally advanced cervical cancer patients underwent pre- and per-treatment 18F-FDG PET (respectively PET1 and PET2). Tumor metabolism was characterized through several delineations using different thresholds, based on a percentage of the maximum uptake, and applied by region-growing. The SUV distribution in PET1 and PET2 within each segmented region was characterized through 7 intensity and histogram-based parameters, 9 shape descriptors and 16 textural features for a total of 1026 parameters. Predictive capability of the extracted parameters was assessed using the area under the receiver operating curve (AUC) associated to univariate logistic regression models and random forest (RF) classifier. In univariate analyses, 36 parameters were highly significant predictors of 3-year OR (p<;0.01), AUC ranging from 0.72 to 0.83. With RF, the Out-of-Bag (OOB) error rate using the totality of the extracted parameters was 26.42% (AUC=0.72). By recursively eliminating the less important variables, OOB error rate of the RF classifier using the nine most important parameters was 13.21% (AUC=0.90). Results suggest that both pre- and per-treatment 18F-FDG PET exams provide meaningful information to predict the tumor recurrence. RF classifier is able to handle a very large number of extracted features and allows the combination of the most prognostic parameters to improve the prediction.

Weighted quantification of ¹8F-FDG tumor metabolism activity using fuzzy-thresholding to predict post-treatment tumor recurrence.

Cervical cancer is one of the most common cancer to affect women worldwide. Despite the efficiency of radiotherapy treatment, some patients present post-treatment tumor recurrence which increases the risk of death. Early outcome prediction could help oncologists to adapt the treatment. Several studies suggest that quantification of tumor activity using (18)FFDG PET imaging could be used to predict post-treatment tumor recurrence. In this paper we study the predictive value of weighted quantification of tumor metabolism extracted by fuzzy-thresholding for tumor recurrence of locally advanced cervical cancer. Fifty-three patients with locally advanced cervical cancer treated by chemo-radiotherapy were considered in our study. For each patient, a coregistered (18)F-FDG PET/CT scan was acquired before the treatment and was segmented using different hard and fuzzy segmentations methods. The tumor activity was extracted through the total lesion glycolysis and through a weighted analog of the total lesion glycolysis using the probability maps provided by the fuzzy segmentations. Outcomes prediction was performed using the area under the receiver operating characteristic curve (AUC) and the Harrell's C-index. Results suggest that weighted quantification of tumor activity seems to be strongly informative and could be used to predict post-treatment tumor recurrence in cervical cancer.