Correlation Between SUV_max and CT Radiomic Analysis Using Lymph Node Density in PET/CT-Based Lymph Node Staging

Abstract

In patients with lung cancer (LC), malignant melanoma (MM), gastroenteropancreatic neuroendocrine tumors (GEP NETs), and prostate cancer (PCA), lymph node (LN) staging is often performed by 18 F-FDG PET/CT (LC and MM), 68 Ga-DOTATOC PET/CT (GEP NET), and 68 Ga-labeled prostate-specific membrane antigen PET/CT (PCA) but is sometimes not accurate because of indeterminate PET findings. To better evaluate malignant LN infiltration, additional surrogate parameters, especially in cases with indeterminate PET findings, would be helpful. The purpose of this study was to evaluate whether SUV max in the PET examination might correlate with semiautomated density measurements of LNs in the CT component of the PET/CT examination. Methods: After approval by the institutional review board, 1,022 LNs in the PET/CT examinations of 148 patients were retrospectively analyzed (LC: 327 LNs of 40 patients; MM: 224 LNs of 33 patients; GEP NET: 217 LNs of 35 patients; and PCA: 254 LNs of 40 patients). PET/CT was performed before surgery, biopsy, chemotherapy, or internal or external radiation therapy, according to the clinical schedule; patients with prior chemotherapy or radiation therapy were excluded. SUV max analyses were based on uptake 60 min after tracer injection, and volumetric CT histogram analyses were based on the unenhanced CT images of the PET/CT scan. Results: PET findings were considered positive or negative on the basis of SUV max in the LN compared with that in the blood pool; histologic confirmation was not available. Of the 1,022 LNs, 331 were PET-positive (3 times the SUV max of the blood pool), 86 were PET-indeterminate (1-3 times the SUV max of the blood pool), and 605 were PET-negative (less than the SUV max of the blood pool). PET-positive LNs had significantly higher CT densities than PET-negative LNs, irrespective of the type of cancer. Conclusion: CT density measurements of LNs in patients with LC, MM, GEP NET, and PCA correlated with 18 F-FDG uptake, 68 Ga-DOTATOC uptake, and 68 Ga-PSMA uptake, respectively, and might therefore serve as an additional surrogate parameter for differentiating between malignant and benign LNs. The use of a 7.5-Hounsfield unit CT density threshold to differentiate between malignant and benign LN infiltration and 20 Hounsfield units to exclude benign LN processes might be possible in clinical routine and would be especially helpful for PET-indeterminate LNs. Oneoft he most widespread and reliable noninvasive staging tools for cancer is PET/CT-namely, 18 F-FDG PET/CT for lung cancer (LC (1,2)) and malignant melanoma (MM (1,3)), 68 Ga-labeled prostate-specific membrane antigen (PSMA) PET/ CT for prostate cancer (PCA (4)), and 68 Ga-DOTATOC PET/ CT for gastroenteropancreatic neuroendocrine tumors (GEP NETs (5)). An exact evaluation of lymph node (LN) status before therapy is crucial to therapy planning. False-positive PET findings are not uncommon with 18 F-FDG PET/CT, because the scan can mistakenly pick up on inflammation due to infectious etiologies, such as in the lung or the h...

Impact of FDG-PET on the Detection of Patients with Lung Cancer at High Risk for ILD

“…18 F-FDG-PET/CT: 18 F-FDG-PET/CT examinations were performed as part of preoperative staging according to clinical routine using a Biograph 6 PET/CT Scanner (Siemens, Erlangen, Germany). Imaging sequences and protocols were described earlier in detail (20,21): after fasting for at least 6 hours (blood glucose level below 150 mg/dl), 4 MBq/kg body weight of 18 F-FDG were administered intravenously 60±5 minutes before the FDG-PET/CT scan. The following examination protocols and reconstruction parameters were applied: slice thickness/reconstruction increment of 5.0/2.5 mm, standard soft-tissue reconstruction kernel B30 and B70 high-resolution.…”

Personalizing Medicine Through Hybrid Imaging and Medical Big Data Analysis

“…In one study, PET, CT, and PET/CT features were used to predicting local tumor control in head and neck cancer [129] by multivariate cox regression with a confidence interval (CI) CI CT and CI PET/CT of 0.73, however, CT-based radiomics overestimated the probability of tumor control in the poor prognostic groups. Another study found a correlation of PET and CT features using lymph node density [176] and concluded that CT density measurements together with PET uptake analysis increases the differentiation between malignant and benign LN. Disease-free survival prediction in non-small cell lung cancer patients can be performed in PET/CT images with an area under the receiver operator characteristic curve (AUC) of 0.68 when employing combined PET/CT features [177].…”

Novel cancer therapies for advanced cutaneous melanoma: The added value of radiomics in the decision making process–A systematic review

“…This might be due to the lower resolution and contrast of nuclear medicine images. 33 In the study by Giesel et al, 41 only LNs were evaluated in order to discriminate those with metastatic infiltration from benign ones. Despite the good results reported, the robustness of radiomic features extracted from small volumes of LNs was overlooked.…”