Potential of ChatGPT and GPT-4 for Data Mining of Free-Text CT Reports on Lung Cancer.

Background The latest large language models (LLMs) solve unseen problems via user-defined text prompts without the need for retraining, offering potentially more efficient information extraction from free-text medical records than manual annotation. Purpose To compare the performance of the LLMs ChatGPT and GPT-4 in data mining and labeling oncologic phenotypes from free-text CT reports on lung cancer by using user-defined prompts. Materials and Methods This retrospective study included patients who underwent lung cancer follow-up CT between September 2021 and March 2023. A subset of 25 reports was reserved for prompt engineering to instruct the LLMs in extracting lesion diameters, labeling metastatic disease, and assessing oncologic progression. This output was fed into a rule-based natural language processing pipeline to match ground truth annotations from four radiologists and derive performance metrics. The oncologic reasoning of LLMs was rated on a five-point Likert scale for factual correctness and accuracy. The occurrence of confabulations was recorded. Statistical analyses included Wilcoxon signed rank and McNemar tests. Results On 424 CT reports from 424 patients (mean age, 65 years ± 11 [SD]; 265 male), GPT-4 outperformed ChatGPT in extracting lesion parameters (98.6% vs 84.0%, P < .001), resulting in 96% correctly mined reports (vs 67% for ChatGPT, P < .001). GPT-4 achieved higher accuracy in identification of metastatic disease (98.1% [95% CI: 97.7, 98.5] vs 90.3% [95% CI: 89.4, 91.0]) and higher performance in generating correct labels for oncologic progression (F1 score, 0.96 [95% CI: 0.94, 0.98] vs 0.91 [95% CI: 0.89, 0.94]) (both P < .001). In oncologic reasoning, GPT-4 had higher Likert scale scores for factual correctness (4.3 vs 3.9) and accuracy (4.4 vs 3.3), with a lower rate of confabulation (1.7% vs 13.7%) than ChatGPT (all P < .001). Conclusion When using user-defined prompts, GPT-4 outperformed ChatGPT in extracting oncologic phenotypes from free-text CT reports on lung cancer and demonstrated better oncologic reasoning with fewer confabulations. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Hafezi-Nejad and Trivedi in this issue.

Association of Tumor Volumetry with Postoperative Outcomes for Cervical Paraganglioma.

Objectives: To analyze the association of tumor volume with outcome after surgery for cervical paraganglioma. Materials and Methods: This retrospective study included consecutive patients undergoing surgery for cervical paraganglioma from 2009-2020. Outcomes were 30-day morbidity, mortality, cranial nerve injury, and stroke. Preoperative CT/MRI was used for tumor volumetry. An association between the volume and the outcomes was explored in univariate and multivariable analyses. A receiver operating characteristic (ROC) curve was plotted, and the area under the curve (AUC) was calculated. The study was conducted and reported according to the STROBE statement. Results: Volumetry was successful in 37/47 (78.8%) of included patients. A 30-day morbidity occurred in 13/47 (27.6%) patients with no mortality. Fifteen cranial nerve lesions occurred in eleven patients. The mean tumor volume was 6.92 cm3 in patients without and 15.89 cm3 in patients with complications (p = 0.035) and 7.64 cm3 in patients without and 16.28 cm3 in patients with cranial nerve injury (p = 0.05). Neither the volume nor Shamblin grade was significantly associated with complications on multivariable analysis. The AUC was 0.691, indicating a poor to fair performance of volumetry in predicting postoperative complications. Conclusions: Surgery for cervical paraganglioma bears a relevant morbidity with a particular risk of cranial nerve lesions. Tumor volume is associated with morbidity, and MRI/CT volumetry can be used for risk stratification.

Deep Learning-based Assessment of Oncologic Outcomes from Natural Language Processing of Structured Radiology Reports.

Purpose: To train a deep natural language processing (NLP) model, using data mined structured oncology reports (SOR), for rapid tumor response category (TRC) classification from free-text oncology reports (FTOR) and to compare its performance with human readers and conventional NLP algorithms. Materials and Methods: In this retrospective study, databases of three independent radiology departments were queried for SOR and FTOR dated from March 2018 to August 2021. An automated data mining and curation pipeline was developed to extract Response Evaluation Criteria in Solid Tumors-related TRCs for SOR for ground truth definition. The deep NLP bidirectional encoder representations from transformers (BERT) model and three feature-rich algorithms were trained on SOR to predict TRCs in FTOR. Models' F1 scores were compared against scores of radiologists, medical students, and radiology technologist students. Lexical and semantic analyses were conducted to investigate human and model performance on FTOR. Results: Oncologic findings and TRCs were accurately mined from 9653 of 12 833 (75.2%) queried SOR, yielding oncology reports from 10 455 patients (mean age, 60 years ± 14 [SD]; 5303 women) who met inclusion criteria. On 802 FTOR in the test set, BERT achieved better TRC classification results (F1, 0.70; 95% CI: 0.68, 0.73) than the best-performing reference linear support vector classifier (F1, 0.63; 95% CI: 0.61, 0.66) and technologist students (F1, 0.65; 95% CI: 0.63, 0.67), had similar performance to medical students (F1, 0.73; 95% CI: 0.72, 0.75), but was inferior to radiologists (F1, 0.79; 95% CI: 0.78, 0.81). Lexical complexity and semantic ambiguities in FTOR influenced human and model performance, revealing maximum F1 score drops of -0.17 and -0.19, respectively. Conclusion: The developed deep NLP model reached the performance level of medical students but not radiologists in curating oncologic outcomes from radiology FTOR.Keywords: Neural Networks, Computer Applications-Detection/Diagnosis, Oncology, Research Design, Staging, Tumor Response, Comparative Studies, Decision Analysis, Experimental Investigations, Observer Performance, Outcomes Analysis Supplemental material is available for this article. © RSNA, 2022.

NMDA Receptor Activation Underlies the Loss of Spinal Dorsal Horn Neurons and the Transition to Persistent Pain after Peripheral Nerve Injury.

Peripheral nerve lesions provoke apoptosis in the dorsal horn of the spinal cord. The cause of cell death, the involvement of neurons, and the relevance for the processing of somatosensory information are controversial. Here, we demonstrate in a mouse model of sciatic nerve injury that glutamate-induced neurodegeneration and loss of γ-aminobutyric acid (GABA)ergic interneurons in the superficial dorsal horn promote the transition from acute to chronic neuropathic pain. Conditional deletion of Grin1, the essential subunit of N-methyl-d-aspartate-type glutamate receptors (NMDARs), protects dorsal horn neurons from excitotoxicity and preserves GABAergic inhibition. Mice deficient in functional NMDARs exhibit normal nociceptive responses and acute pain after nerve injury, but this initial increase in pain sensitivity is reversible. Eliminating NMDARs fully prevents persistent pain-like behavior. Reduced pain in mice lacking proapoptotic Bax confirmed the significance of neurodegeneration. We conclude that NMDAR-mediated neuron death contributes to the development of chronic neuropathic pain.

Stereotaxic injection of a viral vector for conditional gene manipulation in the mouse spinal cord.

Intraparenchymal injection of a viral vector enables conditional gene manipulation in distinct populations of neurons or particular regions of the central nervous system. We demonstrate a stereotaxic injection technique that allows targeted gene expression or silencing in the dorsal horn of the mouse spinal cord. The surgical procedure is brief. It requires laminectomy of a single vertebra, providing for quick recovery of the animal and unimpaired motility of the spine. Controlled injection of a small vector suspension volume at low speed and use of a microsyringe with beveled glass cannula minimize the tissue lesion. The local immune response to the vector depends on the intrinsic properties of the virus employed; in our experience, it is minor and short-lived when a recombinant adeno-associated virus is used. A reporter gene such as enhanced green fluorescent protein facilitates monitoring spatial distribution of the vector, and the efficacy and cellular specificity of the transfection.