PET-F2I: A Comprehensive Benchmark and Parameter-Efficient Fine-Tuning of LLMs for PET/CT Report Impression Generation

• URL: http://arxiv.org/abs/2603.10560v1
• Date: Wed, 11 Mar 2026 09:08:25 GMT
• Title: PET-F2I: A Comprehensive Benchmark and Parameter-Efficient Fine-Tuning of LLMs for PET/CT Report Impression Generation
• Abstract summary: We introduce PET-F2I-41K (PET Findings-to-Impression Benchmark), a large-scale benchmark for PET/CT impression generation using LLMs.<n>Using PET-F2I-41K, we conduct a comprehensive evaluation of 27 models across proprietary frontier LLMs, open-source generalist models, and medical-domain LLMs.<n>PET-F2I-7B achieves substantial gains and a 3.0x improvement in entity coverage over the strongest baseline.
• Score: 25.57267405761328
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: PET/CT imaging is pivotal in oncology and nuclear medicine, yet summarizing complex findings into precise diagnostic impressions is labor-intensive. While LLMs have shown promise in medical text generation, their capability in the highly specialized domain of PET/CT remains underexplored. We introduce PET-F2I-41K (PET Findings-to-Impression Benchmark), a large-scale benchmark for PET/CT impression generation using LLMs, constructed from over 41k real-world reports. Using PET-F2I-41K, we conduct a comprehensive evaluation of 27 models across proprietary frontier LLMs, open-source generalist models, and medical-domain LLMs, and we develop a domain-adapted 7B model (PET-F2I-7B) fine-tuned from Qwen2.5-7B-Instruct via LoRA. Beyond standard NLG metrics (e.g., BLEU-4, ROUGE-L, BERTScore), we propose three clinically grounded metrics - Entity Coverage Rate (ECR), Uncovered Entity Rate (UER), and Factual Consistency Rate (FCR) - to assess diagnostic completeness and factual reliability. Experiments reveal that neither frontier nor medical-domain LLMs perform adequately in zero-shot settings. In contrast, PET-F2I-7B achieves substantial gains (e.g., 0.708 BLEU-4) and a 3.0x improvement in entity coverage over the strongest baseline, while offering advantages in cost, latency, and privacy. Beyond this modeling contribution, PET-F2I-41K establishes a standardized evaluation framework to accelerate the development of reliable and clinically deployable reporting systems for PET/CT.

Related papers

• The autoPET3 Challenge: Automated Lesion Segmentation in Whole-Body PET/CT $\unicode{x2013}$ Multitracer Multicenter Generalization [67.92281859568168]
  We report the design and results of the third autoPET challenge (MICCAI 2024)<n>It benchmarked automated lesion segmentation in whole-body PET/CT under a compositional generalization setting.<n>The top-ranked algorithm achieved a mean DSC of 0.66, FNV of 3.18 mL, and FPV of 2.78 mL across all four test conditions.
  arXiv  Detail & Related papers  (2026-05-07T07:07:09Z)
• Cross-Modal Knowledge Distillation for PET-Free Amyloid-Beta Detection from MRI [8.083880525992651]
  A$$ positivity is crucial for early diagnosis of Alzheimer's disease but typically requires PET imaging.<n>We propose a PET-guided knowledge distillation framework that enables A$$ prediction from MRI alone.
  arXiv  Detail & Related papers  (2026-04-14T10:52:55Z)
• A Federated and Parameter-Efficient Framework for Large Language Model Training in Medicine [59.78991974851707]
  Large language models (LLMs) have demonstrated strong performance on medical benchmarks, including question answering and diagnosis.<n>Most medical LLMs are trained on data from a single institution, which faces limitations in generalizability and safety in heterogeneous systems.<n>We introduce the model-agnostic and parameter-efficient federated learning framework for adapting LLMs to medical applications.
  arXiv  Detail & Related papers  (2026-01-29T18:48:21Z)
• Unveiling and Bridging the Functional Perception Gap in MLLMs: Atomic Visual Alignment and Hierarchical Evaluation via PET-Bench [48.60251555171943]
  Multimodal Large Language Models (MLLMs) have demonstrated remarkable proficiency in tasks such as abnormality detection and report generation for anatomical modalities.<n>In this work, we quantify a fundamental functional perception gap: the inability of current vision encoders to decode functional tracer biodistribution independent of morphological priors.<n>We introduce PET-Bench, the first large-scale functional imaging benchmark comprising 52,308 hierarchical QA pairs from 9,732 multi-site, multi-tracer PET studies.<n>Our results demonstrate that AVA effectively bridges the perception gap, transforming CoT from a source of hallucination into a robust inference tool and improving diagnostic
  arXiv  Detail & Related papers  (2026-01-06T05:58:50Z)
• Vision-Language Models for Automated 3D PET/CT Report Generation [13.781844347232079]
  AutoPET/CT report generation is increasingly important for reducing clinical workload.<n>PETRG-3D is an end-to-end 3D dual-branch framework that encodes PET and CT volumes and incorporates style-adaptive prompts.<n>PETRG-Score is a lymphoma-specific evaluation protocol that measures metabolic and structural findings across curated anatomical regions.
  arXiv  Detail & Related papers  (2025-11-25T10:07:57Z)
• Diffusion Bridge Networks Simulate Clinical-grade PET from MRI for Dementia Diagnostics [20.93414282937884]
  We present SiM2P, a framework that learns a probabilistic mapping from MRI and auxiliary patient information to simulate FDG-PET images of diagnostic quality.<n>SiM2P significantly improved the overall diagnostic accuracy of differentiating between three groups from 75.0% to 84.7%.
  arXiv  Detail & Related papers  (2025-10-17T11:42:11Z)
• HiDe-PET: Continual Learning via Hierarchical Decomposition of Parameter-Efficient Tuning [55.88910947643436]
  We propose a unified framework for continual learning (CL) with pre-trained models (PTMs) and parameter-efficient tuning (PET)<n>We present Hierarchical Decomposition PET (HiDe-PET), an innovative approach that explicitly optimize the objective through incorporating task-specific and task-shared knowledge.<n>Our approach demonstrates remarkably superior performance over a broad spectrum of recent strong baselines.
  arXiv  Detail & Related papers  (2024-07-07T01:50:25Z)
• Contrastive Diffusion Model with Auxiliary Guidance for Coarse-to-Fine PET Reconstruction [62.29541106695824]
  This paper presents a coarse-to-fine PET reconstruction framework that consists of a coarse prediction module (CPM) and an iterative refinement module (IRM) By delegating most of the computational overhead to the CPM, the overall sampling speed of our method can be significantly improved. Two additional strategies, i.e., an auxiliary guidance strategy and a contrastive diffusion strategy, are proposed and integrated into the reconstruction process.
  arXiv  Detail & Related papers  (2023-08-20T04:10:36Z)
• CA-LoRA: Adapting Existing LoRA for Compressed LLMs to Enable Efficient Multi-Tasking on Personal Devices [78.16679232748196]
  We introduce a Compression-Aware LoRA (CA-LoRA) framework to transfer Large Language Models (LLMs) to other tasks. Experiment results demonstrate that CA-LoRA outperforms the vanilla LoRA methods applied to a compressed LLM. The source code of CA-LoRA is available at https://github.com/thunlp/CA-LoRA.
  arXiv  Detail & Related papers  (2023-07-15T04:37:11Z)
• A Unified Continual Learning Framework with General Parameter-Efficient Tuning [56.250772378174446]
  "Pre-training $rightarrow$ downstream adaptation" presents both new opportunities and challenges for Continual Learning. We position prompting as one instantiation of PET, and propose a unified CL framework, dubbed as Learning-Accumulation-Ensemble (LAE) PET, e.g., using Adapter, LoRA, or Prefix, can adapt a pre-trained model to downstream tasks with fewer parameters and resources.
  arXiv  Detail & Related papers  (2023-03-17T15:52:45Z)
• Synthetic PET via Domain Translation of 3D MRI [1.0052333944678682]
  We use a dataset of 56 $18$F-FDG-PET/MRI exams to train a 3D residual UNet to predict physiologic PET uptake from whole-body T1-weighted MRI. The predicted PET images are forward projected to produce synthetic PET time-of-flight sinograms that can be used with vendor-provided PET reconstruction algorithms.
  arXiv  Detail & Related papers  (2022-06-11T21:32:40Z)
• A resource-efficient deep learning framework for low-dose brain PET image reconstruction and analysis [13.713286047709982]
  We propose a resource-efficient deep learning framework for L-PET reconstruction and analysis, referred to as transGAN-SDAM. The transGAN generates higher quality F-PET images, and then the SDAM integrates the spatial information of a sequence of generated F-PET slices to synthesize whole-brain F-PET images.
  arXiv  Detail & Related papers  (2022-02-14T08:40:19Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.