Dataset Information

An [18F]FDG-PET/CT deep learning method for fully automated detection of pathological mediastinal lymph nodes in lung cancer patients.

ABSTRACT:

Purpose

The identification of pathological mediastinal lymph nodes is an important step in the staging of lung cancer, with the presence of metastases significantly affecting survival rates. Nodes are currently identified by a physician, but this process is time-consuming and prone to errors. In this paper, we investigate the use of artificial intelligence-based methods to increase the accuracy and consistency of this process.

Methods

Whole-body ¹⁸F-labelled fluoro-2-deoxyglucose ([18F]FDG) positron emission tomography/computed tomography ([18F]FDG-PET/CT) scans (Philips Gemini TF) from 134 patients were retrospectively analysed. The thorax was automatically located, and then slices were fed into a U-Net to identify candidate regions. These regions were split into overlapping 3D cubes, which were individually predicted as positive or negative using a 3D CNN. From these predictions, pathological mediastinal nodes could be identified. A second cohort of 71 patients was then acquired from a different, newer scanner (GE Discovery MI), and the performance of the model on this dataset was tested with and without transfer learning.

Results

On the test set from the first scanner, our model achieved a sensitivity of 0.87 (95% confidence intervals [0.74, 0.94]) with 0.41 [0.22, 0.71] false positives/patient. This was comparable to the performance of an expert. Without transfer learning, on the test set from the second scanner, the corresponding results were 0.53 [0.35, 0.70] and 0.24 [0.10, 0.49], respectively. With transfer learning, these metrics were 0.88 [0.73, 0.97] and 0.69 [0.43, 1.04], respectively.

Conclusion

Model performance was comparable to that of an expert on data from the same scanner. With transfer learning, the model can be applied to data from a different scanner. To our knowledge it is the first study of its kind to go directly from whole-body [18F]FDG-PET/CT scans to pathological mediastinal lymph node localisation.

SUBMITTER: Wallis D

PROVIDER: S-EPMC8803782 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:This study assessed whether preoperative maximum standardized uptake value (SUVmax) of metastatic lymph nodes (LNs) measured by F-fluoro-2-deoxy-D-glucose positron emission tomography/computed tomography (F-FDG PET/CT) could improve the prediction of prognosis in gastric cancer.One hundred fifty-one patients with gastric cancer and pathologically confirmed LN involvement who had undergone preoperative F-FDG PET/CT prior to curative surgical resection were retrospectively enrolled. To obtain nodal SUVmax, a transaxial image representing the highest F-FDG uptake was carefully selected, and a region of interest was manually drawn on the highest F-FDG accumulating LN. Conventional prognostic parameters and PET findings (primary tumor and nodal SUVmax) were analyzed for prediction of recurrence-free survival (RFS) and overall survival (OS). Furthermore, prognostic accuracy of survival models was assessed using c-statistics.Of the 151 patients, 38 (25%) experienced recurrence and 34 (23%) died during follow-up (median follow-up, 48 months; range, 5-74 months). Twenty-seven patients (18%) showed positive F-FDG nodal uptake (range, 2.0-22.6). In these 27 patients, a receiver-operating characteristic curve demonstrated a nodal SUVmax of 2.8 to be the optimal cutoff for predicting RFS and OS. The univariate and multivariate analyses showed that nodal SUVmax (hazard ratio [HR]?= 2.71, P < 0.0001), pathologic N (pN) stage (HR = 2.58, P = 0.0058), and pathologic T (pT) stage (HR = 1.77, P = 0.0191) were independent prognostic factors for RFS. Also, nodal SUVmax (HR = 2.80, P < 0.0001) and pN stage (HR = 2.28, P = 0.0222) were independent prognostic factors for OS. A predictive survival model incorporating conventional risk factors (pT/pN stage) gave a c-statistic of 0.833 for RFS and 0.827 for OS, whereas a model combination of nodal SUVmax with pT/pN stage gave a c-statistic of 0.871 for RFS (P = 0.0355) and 0.877 for OS (P = 0.0313).Nodal SUVmax measured by preoperative F-FDG PET/CT is an independent prognostic factor for RFS and OS. Combining nodal SUVmax with pT/pN staging can improve survival prediction precision in patients with gastric cancer.