Anatomy-Aware Promptable Segmentation with Online Interactive Training for AUTOPET V

Pablo Lozano-Jimenez; Sergio Romero-Tapiador; Ruben Tolosana

cs.CVAug 28, 2026

Anatomy-Aware Promptable Segmentation with Online Interactive Training for AUTOPET V

Authors: Pablo Lozano-Jimenez, Sergio Romero-Tapiador, Ruben Tolosana

Organizations: University of Amsterdam, Faculty of Science, Amsterdam, The Netherlands · BiometricsAI, Universidad Autónoma de Madrid, 28049 Madrid, Spain

Abstract

We present an anatomy-aware, promptable model for whole-body lesion segmentation in FDG and PSMA PET/CT, developed for the AUTOPET V challenge. The proposed method is built as family of nnU-Net-based models and trained in two stages: i) a pre-training stage that produces a strong initial segmentation, and ii) an online interactive stage that learns to exploit scribble prompts, refining the prediction over successive interactions. Anatomical context is incorporated through organ supervision using a single shared head that predicts lesions and organs from the same features, which reduces false positives arising from physiological uptake. Also as the tracer (i.e., FDG/PSMA) is not provided at inference, we add a tracer classifier based on image processing and a random forest over coronal MIP features, routing each study to a combined FDG+PSMA model or to a PSMA-specific model. Across four-fold cross-validation the organ-supervised model achieves the best and most stable performance, the interactive stage improves the Dice score monotonically with each prompt, and PSMA-specific training yields the strongest tracer-wise results.

Figures & tables

		V1	V2	V3	V4	V5
Overall	DICE	$0.449{\pm}0.024$	$0.507{\pm}0.017$	$0.551{\pm}0.025$	$0.557{\pm}0.067$	$\mathbf{0.578{\pm}0.019}$
	DMM	$0.752{\pm}0.013$	$0.767{\pm}0.014$	$0.785{\pm}0.019$	$0.776{\pm}0.019$	$\mathbf{0.794{\pm}0.020}$
	Metric	$0.600{\pm}0.019$	$0.637{\pm}0.015$	$0.668{\pm}0.022$	$0.666{\pm}0.043$	$\mathbf{0.686{\pm}0.020}$
FDG	DICE	$0.415{\pm}0.028$	$0.498{\pm}0.018$	$0.559{\pm}0.032$	$\mathbf{0.573{\pm}0.093}$	$-$
	DMM	$0.606{\pm}0.028$	$0.629{\pm}0.009$	$\mathbf{0.670{\pm}0.025}$	$0.654{\pm}0.029$	$-$
	Metric	$0.511{\pm}0.028$	$0.563{\pm}0.014$	$\mathbf{0.614{\pm}0.028}$	$0.613{\pm}0.061$	$-$

Table 1: Phase-1 results. Training and evaluation were done using AUTOPET V Dataset. V1: Baseline; V2: Architecture; V3: Organ supervision; V4: Combined model; V5: Tracer classifier & PSMA-specific training

			w/o	S1	S2	S3	S4	S5	AUC
$\sigma=5$	FDG	DICE	$0.585$	$0.708$	$0.803$	$0.847$	$0.875$	$0.885$	$0.784$
		DMM	$0.681$	$0.708$	$0.717$	$0.725$	$0.734$	$0.740$	$0.717$
		Metric	$0.633$	$0.708$	$0.760$	$0.786$	$0.804$	$0.813$	$0.751$
	PSMA	DICE	$0.581$	$0.703$	$0.740$	$0.762$	$0.778$	$0.788$	$\mathbf{0.725}$
		DMM	$0.863$	$0.869$	$0.871$	$0.872$	$0.872$	$0.874$	$\mathbf{0.870}$
		Metric	$0.722$	$0.786$	$0.806$	$0.817$	$0.825$	$0.831$	$\mathbf{0.798}$

Table 2: Phase-2 results with 1-fold training and evaluation on AUTOPET V dataset

Team name	algorithm name (as submitted on grand-challenge)	data pre-processing	data post-processing	training data augmentation
UAM_team	UAM_team _submission	nnUNet Preprocessing	-	Random Brightness, Random Gamma, Random Rotation

Table 3: Algorithm details

Explore similar work

Pretrained, Curriculum-Tuned, and Ensembled: A Tracer-Aware Interactive Segmentation Pipeline for AutoPET V
Aug 31, 2026Xinglong Liang, Chunyao Lu, Tianyu Zhang +4Interactive Medical Image Segmentation PET/CT Imaging
Department of Radiology, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX, Amsterdam, the Netherlands · Department of Radiology and Nuclear Medicine, Radboud University Medical Centre, Geert Grooteplein 10, 6525 GA, Nijmegen, The Netherlands · Faculty of Applied Sciences, Macao Polytechnic University, 999078, Macao Special Administrative Region of China +2
BS: Take the Hint - Interactive Multitracer PET/CT Lesion Segmentation with a Scribble-Conditioned ResEnc U-Net
Sep 1, 2026Marven Sherif, Amgad Elmasry, Youssef Ghazal +1Interactive Medical Image Segmentation U-Net
Brightskies
The autoPET3 Challenge: Automated Lesion Segmentation in Whole-Body PET/CT $\unicode{x2013}$ Multitracer Multicenter Generalization
May 7, 2026Jakob Dexl, Katharina Jeblick, Andreas Mittermeier +27Domain Generalization PET/CT Imaging
Department of Radiology, LMU University Hospital, LMU Munich, Munich, Germany · Munich Center for Machine Learning (MCML), Munich, Germany · Comprehensive Pneumology Center (CPC-M), Member of the German Center for Lung Research (DZL), Munich, Germany +13

Aug 31, 2026J/K move · Enter open · S save

Pretrained, Curriculum-Tuned, and Ensembled: A Tracer-Aware Interactive Segmentation Pipeline for AutoPET V

Xinglong Liang, Chunyao Lu, Tianyu Zhang, Jiaju Huang+3

Department of Radiology, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX, Amsterdam, the Netherlands · Department of Radiology and Nuclear Medicine, Radboud University Medical Centre, Geert Grooteplein 10, 6525 GA, Nijmegen, The Netherlands · Faculty of Applied Sciences, Macao Polytechnic University, 999078, Macao Special Administrative Region of China+2

Interactive Medical Image Segmentation PET/CT Imaging Medical Image Segmentation PET/CT Tumor Segmentation

Open paper AskPDF arXiv

Interactive lesion segmentation in whole-body PET/CT requires a model to provide a strong initial prediction while also responding efficiently to sparse corrective scribbles during inference. This setting is particularly challenging because tracer distributions, physiological uptake patterns, lesion appearance, and acquisition characteristics differ substantially between FDG and PSMA studies. We present TRIAGE, Tracer-aware Refinement via Interactive Anatomy-Guided sEgmentation. The core backbone is a 3D STU-Net initialized through masked autoencoding pre-training with an asynchronous masking strategy, aiming to learn transferable anatomical and cross-modal representations before task-specific fine-tuning. In parallel, we train an auxiliary organ segmentation model whose predictions provide explicit anatomical context and help distinguish physiological uptake from malignant lesions. A dedicated tracer classifier first routes each study to an FDG- or PSMA-specific branch. Within each branch, a first-stage segmentation model consumes CT, PET, and organ context to generate an initial lesion mask. The initial prediction is then combined with cumulative foreground/background scribbles and refined by a second interactive segmentation network. The FDG and PSMA branches share the same overall processing pipeline but are trained independently to account for tracer-specific appearance and error modes. We additionally employ curriculum-style training and model ensembling to improve robustness across interaction steps and heterogeneous cohorts. Experiments are conducted using the official AutoPET V data and ten-fold split; quantitative results, ablations, and final test-set performance are left as placeholders to be completed after the challenge evaluation. Code: https://github.com/Liiiii2101/AUTOPET2026-MEDAI.

Abstract

Explore similar work

Pretrained, Curriculum-Tuned, and Ensembled: A Tracer-Aware Interactive Segmentation Pipeline for AutoPET V

BS: Take the Hint - Interactive Multitracer PET/CT Lesion Segmentation with a Scribble-Conditioned ResEnc U-Net

The autoPET3 Challenge: Automated Lesion Segmentation in Whole-Body PET/CT \unicodex2013\unicode{x2013}\unicodex2013 Multitracer Multicenter Generalization

The autoPET3 Challenge: Automated Lesion Segmentation in Whole-Body PET/CT $\unicode{x2013}$ Multitracer Multicenter Generalization