AI summaryⓘ
The authors tackle the problem of segmenting lesions in whole-body PET/CT scans, which is tricky because lesions and normal tissue look different across scans and tracers. They created an interactive system where users provide scribbles to label parts as either lesion (foreground) or non-lesion (background), helping the model learn better. Their model builds on a known U-Net architecture and uses these scribbles along with PET and CT images to improve accuracy. Initially, the model performs moderately without scribbles but significantly improves after a few rounds of user corrections. This shows that interactive hints can greatly enhance automated lesion segmentation even when the model alone struggles.
PET/CTlesion segmentationU-Netinteractive segmentationscribble annotationresidual encoderautoPET challengeDice scoreF1 score
Authors
Marven Sherif, Amgad Elmasry, Youssef Ghazal, Ayman Elghotni
Abstract
Automated lesion segmentation in whole-body PET/CT is complicated by the variety of physiological tracer uptake patterns and by the differing appearance of lesions across tracers. The autoPET/CT V challenge addresses this by making segmentation interactive: user scribbles marking foreground and background are supplied alongside the image, and the algorithm is expected to exploit them. We present our submission, a scribble-conditioned residual encoder U-Net operating on four input channels: CT, PET, and a sparse scribble map for each of foreground and background. The network is initialised from the autoPET-III winning weights and extended from two to four input channels, with the two scribble channels zero-initialised so that the pretrained representation is preserved exactly at initialisation. Every model is fine-tuned per fold from the corresponding autoPET-III fold checkpoint, so that no validation case is seen during pretraining. PET intensities are normalised against a per-scan aorta blood-pool reference derived from a CT segmentation, which removes tracer- and centre-specific scaling without requiring lesion labels. At inference the five fold models are ensembled by averaging their softmax outputs per sliding-window patch, before Gaussian-weighted stitching. On the challenge's five-fold split, with each fold evaluated on its own validation cases, mean Dice is 0.554 and mean lesion-level F1 is 0.528 without scribbles, rising to 0.751 and 0.733 after five correction rounds. About 85% of that gain follows the first scribble, and the spread between fold models narrows five-fold over the same rounds, so interaction largely compensates for how well or badly a given model segments unaided.