Shuai Liu1,2,3,4, Zheng Feng2,5, Jiajia Zhang2,5,6, Huijuan Ge2,7, Xiaohua Wu2,5, Shaoli Song1,2,3,4. 1. Department of Nuclear Medicine, Fudan University Shanghai Cancer Center, Shanghai, China. 2. Department of Oncology, Shanghai Medical College, Fudan University, Shanghai, China. 3. Center for Biomedical Imaging, Fudan University, Shanghai, China. 4. Shanghai Engineering Research Center of Molecular Imaging Probes, Shanghai, China. 5. Department of Gynecological Oncology, Fudan University Shanghai Cancer Center, Shanghai, China. 6. Department of Nursing, Fudan University Shanghai Cancer Center, Shanghai, China. 7. Department of Pathology, Fudan University Shanghai Cancer Center, Shanghai, China.
Abstract
BACKGROUND: We aimed to establish an effective 2-deoxy-2-fluorodeoxyglucose (18F-FDG) positron emission tomography/computed tomography (PET/CT) based nomogram for pelvic lymph node (PLN) metastasis prediction in early-stage uterine cervical squamous cell cancer. METHODS: A predictive model was developed in a cohort that consisted of 351 patients with stage IB-IIA [International Federation of Gynecology and Obstetrics (FIGO) 2009] uterine cervical squamous cell cancer. All patients underwent a preoperative PET/CT scan and subsequent radical surgery between 2010 and 2017, with 241 and 110 patients allotted into training and external validation cohorts. The chi-square (χ2) test and the logistic regression analysis were used to compare the clinical and PET/CT parameters with PLN metastasis. A nomogram was developed and validated by internal and external validation. RESULTS: In the training cohort, 82 (34.0%) patients had positive PLNs identified in the preoperative PET/CT scan. Among them, 46 (56.1%) were pathologically confirmed. There were 30 (18.9%) PET/CT scan-negative patients found to have PLN metastasis. The χ2 test and logistic regression showed that only the squamous cell carcinoma antigen (SCCA) level (P=0.039) and maximum standardized uptake value (SUVmax) of PLN (nSUVmax, P=0.001) were independent predictors for PLN metastasis. A predictive nomogram based on these 2 parameters was developed with a C-index [95% confidence interval (CI)] of 0.854 (0.772-0.937) on internal validation and 0.836 (0.723-0.948) on the external validation. Compared to nSUVmax alone, our nomogram showed elevated sensitivity (70.5%, 73.1% vs. 60.5%), specificity (94.4%, 86.4% vs. 78.2%), and positive predictive value (PPV) (93.9%, 86.4% vs. 56.1%) in both the training and validation cohorts. CONCLUSIONS: We successfully developed a noninvasive and convenient nomogram for preoperative identification of PLN metastasis in early-stage squamous cell cervical cancer. 2021 Quantitative Imaging in Medicine and Surgery. All rights reserved.
BACKGROUND: We aimed to establish an effective 2-deoxy-2-fluorodeoxyglucose (18F-FDG) positron emission tomography/computed tomography (PET/CT) based nomogram for pelvic lymph node (PLN) metastasis prediction in early-stage uterine cervical squamous cell cancer. METHODS: A predictive model was developed in a cohort that consisted of 351 patients with stage IB-IIA [International Federation of Gynecology and Obstetrics (FIGO) 2009] uterine cervical squamous cell cancer. All patients underwent a preoperative PET/CT scan and subsequent radical surgery between 2010 and 2017, with 241 and 110 patients allotted into training and external validation cohorts. The chi-square (χ2) test and the logistic regression analysis were used to compare the clinical and PET/CT parameters with PLN metastasis. A nomogram was developed and validated by internal and external validation. RESULTS: In the training cohort, 82 (34.0%) patients had positive PLNs identified in the preoperative PET/CT scan. Among them, 46 (56.1%) were pathologically confirmed. There were 30 (18.9%) PET/CT scan-negative patients found to have PLN metastasis. The χ2 test and logistic regression showed that only the squamous cell carcinoma antigen (SCCA) level (P=0.039) and maximum standardized uptake value (SUVmax) of PLN (nSUVmax, P=0.001) were independent predictors for PLN metastasis. A predictive nomogram based on these 2 parameters was developed with a C-index [95% confidence interval (CI)] of 0.854 (0.772-0.937) on internal validation and 0.836 (0.723-0.948) on the external validation. Compared to nSUVmax alone, our nomogram showed elevated sensitivity (70.5%, 73.1% vs. 60.5%), specificity (94.4%, 86.4% vs. 78.2%), and positive predictive value (PPV) (93.9%, 86.4% vs. 56.1%) in both the training and validation cohorts. CONCLUSIONS: We successfully developed a noninvasive and convenient nomogram for preoperative identification of PLN metastasis in early-stage squamous cell cervical cancer. 2021 Quantitative Imaging in Medicine and Surgery. All rights reserved.
Authors: Laurie L Brunette; Shahram Bonyadlou; Lingyun Ji; Susan Groshen; Dena Shuster; Arjun Mehta; Richard Sposto; Koji Matsuo; Yvonne G Lin; Lynda D Roman Journal: Clin Nucl Med Date: 2018-11 Impact factor: 7.794