Evaluating the Diagnostic Performance of Large Language Models on Complex Multimodal Medical Cases

Large language models showed interpretative reasoning in solving diagnostically challenging medical cases.

Combined 18F-FDG and 11C-acetate positron emission tomography/computed tomography in staging and treatment decision in patients with hepatocellular carcinoma: A cost-effectiveness analysis

e16176

Background: Dual-tracer positron emission tomography/computed tomography (PET/CT) with 18F-fluorodeoxyglucose and 11C-acetate is increasingly used in the management of hepatocellular carcinoma (HCC). However, considering the high cost of dual-tracer PET/CT, there is a need to assess its value by considering both the efficacy and cost. Methods: In this retrospective cohort study, 223 dual-tracer PET/CT scans were performed in 206 patients between 2014 and 2020 for tumor staging for radiologically or histologically confirmed HCC. New lesion detection rate, changes in the Barcelona Clinic Liver Cancer (BCLC) classification, and changes in treatment allocation based on dual-tracer PET/CT were evaluated. Validation was performed on a cohort of staging patients from another institution (n = 48). A cost-benefit analysis model was built for the incurred costs and the impact of dual-tracer PET/CT findings on treatment strategy over conventional imaging was studied. Results: Of the 223 patients that underwent dual-tracer PET/CT for staging, new lesions were detected in 33 (14.8%) patients, resulting in BCLC upstaging in 26 (11.6%) and treatment modification in 16 (7.2%) cases. The external staging cohort showed similar findings where dual-tracer PET/CT resulted in 12.5% (6/48) BCLC upstaging and 10.4% (5/48) treatment modifications (concordance-index: 0.85; 95% CI: 0.82-0.88). In subgroup analysis of 200 patients performed for pre-treatment metastatic screening, dual-tracer PET/CT detected extra-hepatic metastasis in 11, 10, and 16 patients, respectively, and led to treatment modifications in all of these patients. Sixteen locoregional therapies (8%) were avoided (8 resections, 3 transplantations, and 5 TACEs), with an estimated cost saving of US$161,665 (US$808 / patient) from the use of dual-tracer PET/CT. Sensitivity analyses suggested that dual-tracer PET/CT is consistently more cost-saving than conventional imaging, irrespective of the changes in the cost of imaging, procedures, and the duration of hospital stay (range: US$112,945-210,385). The cost of PET/CT had the greatest influence on the cost savings. Conclusions: Comparing to conventional imaging, dual-tracer PET/CT improves tumor staging, alters treatment allocation and is cost-saving in patients undergoing pre-treatment metastatic screening. Our results demonstrated that the use of dual-tracer PET/CT provides additional value and improves patient care.

Sequential trans-arterial chemoembolization and stereotactic body radiotherapy followed by immunotherapy (START-FIT) for locally advanced hepatocellular carcinoma: A single-arm, phase II trial

4091

Background: Previous studies proposed therapeutic synergy between loco-regional therapies and checkpoint inhibitors in hepatocellular carcinoma (HCC). We aimed to study the safety and efficacy of sequential transarterial chemoembolization (TACE) and stereotactic body radiotherapy (SBRT) followed by Avelumab in patients with locally advanced HCC. Methods: Patients with locally advanced HCC not suitable for curative resections were eligible. All patients had HCC ≥ 5cm, tumor nodules ≤3, and child-Pugh A5-B7 liver function. Tumors with distant metastasis, main portal vein (VP4) invasion, or inferior vena cava (VV3) invasion were excluded. Patients underwent single episode of TACE followed by 5-fraction SBRT (28 days afterwards), followed by Avelumab (10mg per kg) 14 days afterwards and every 2 weeks thereafter. The primary endpoint was percentage of patients amendable to curative surgery, defined as R0 resection with sufficient remnant liver volume and function. Secondary endpoints were objective response rate (ORR) per modified Response Evaluation Criteria in Solid Tumors (mRECIST) version 1.1, survivals, and treatment-related adverse event (TRAE). The sample size assumed that around 20% patients amendable to surgery after experimental treatment compared to 5% of historical institutional results after TACE. Modified Simon two-stage optimal design was used (power, 80%; a = 0.05; P0 = 0.05; P1 = 0.20; n1 = 10; n = 29 with an additional four patients to allow for drop-out or other reasons). Results: Out of 67 patients screened, 33 patients were enrolled. The median sum of diameter of lesion(s) was 15.1cm (range: 5.3-31.1cm), and 21 (63.6%) had marcrovascular invasion (n=13, hepatic vein, n=3 branched portal vein, n=5 both). After a median follow-up of 17.2 months (range: 3.5-31.6 months), 3 (9.1%) patients had tumor downstaged with curative surgery done. The objective response rate was 62.5% (95% CI, 45.3-77.1%), of whom 15 had complete response (CR) (43.8%) and 6 had partial response (18.7%). The median overall survival (OS) and progression-free survival was 30.3 months (95% CI: 22.7-37.8 months) and 20.7 months (95% CI: 14.6-26.8 months) respectively. All three patients with surgery done were alive at 2 years (100%). For 15 patients had CR, the 2-year OS rate was 92.9% without surgery. Ten patients (30.3%) experienced ≥ grade 3 TRAEs, commonly transient increase in alanine / aspartate aminotransferase (n=4, 12.1%) and bilirubin (n=2, 6%) level after TACE. Five patients (15.2 %) developed ≥ grade 3 immune-related adverse events. Conclusions: Although merely 9% of patients were downstaged to receive curative surgery, combined locoregional treatment and immunotherapy is safe and resulted in an unexpectedly high CR rate of 43% and median OS of 30 months in patients with locally advanced unresectable HCC. Clinical trial information: NCT 03817736.

Adaptive Contrast for Image Regression in Computer-Aided Disease Assessment

Image regression tasks for medical applications, such as bone mineral density (BMD) estimation and left-ventricular ejection fraction (LVEF) prediction, play an important role in computer-aided disease assessment. Most deep regression methods train the neural network with a single regression loss function like MSE or L1 loss. In this paper, we propose the first contrastive learning framework for deep image regression, namely AdaCon, which consists of a feature learning branch via a novel adaptive-margin contrastive loss and a regression prediction branch. Our method incorporates label distance relationships as part of the learned feature representations, which allows for better performance in downstream regression tasks. Moreover, it can be used as a plug-and-play module to improve performance of existing regression methods. We demonstrate the effectiveness of AdaCon on two medical image regression tasks, i.e., bone mineral density estimation from X-ray images and left-ventricular ejection fraction prediction from echocardiogram videos. AdaCon leads to relative improvements of 3.3% and 5.9% in MAE over state-of-the-art BMD estimation and LVEF prediction methods, respectively.

Presence of tumour capsule on contrast-enhanced CT is associated with improved outcomes of stereotactic body radiation therapy in hepatocellular carcinoma patients

PURPOSE

Stereotactic body radiation therapy (SBRT) is a novel local therapy for the treatment of hepatocellular carcinoma (HCC). While effective, there is currently no reliable radiological marker to guide patient selection. In this study, we investigated the prognostic value of capsule appearance on contrast-enhanced computed tomography (CT) for patients undergoing SBRT.

MATERIALS AND METHODS

Between 2006 and 2017, 156 consecutive patients with Child-Pugh score class A/B and HCC ≥ 5 cm who underwent SBRT were retrospectively analysed. Baseline triple-phase CTs of the abdomen were reviewed for the presence of capsule appearances and correlated with objective response rate (ORR), overall survival (OS) and pattern of treatment failure.

RESULTS

Capsule appearance on CT was present in 83 (53.2%) patients. It was associated with improved ORR by Response Evaluation Criteria in Solid Tumours (RECIST) (60.2 vs. 24.7%, p < 0.001) and Modified Response Evaluation Criteria in Solid Tumours (mRECIST) (78.3 vs. 34.2%, p < 0.001). The presence of a capsule was also associated with superior 2‑year local control (89.1 vs. 51.4%, p < 0.001) and 2‑year OS (34.1 vs. 14.8%, p < 0.01). Hepatic out-field failure was the dominant mode of progression, which was less common in patients with intact capsule (54.2 vs. 60.3%, p = 0.01).

CONCLUSION

Capsule appearance on CT could potentially be a non-invasive prognostic marker for selecting HCC patients to undergo SBRT. A larger cohort is warranted to validate our findings.

Regional liver function analysis with gadoxetic acid-enhanced MRI and virtual hepatectomy: prediction of postoperative short-term outcomes for HCC

OBJECTIVES

To explore the role of quantitative regional liver function assessed by preoperative gadoxetic acid-enhanced MRI with computer-aided virtual hepatectomy to predict short-term outcomes after major hepatectomy for HCC.

METHODS

We retrospectively reviewed the records of 133 consecutive patients with HCC who underwent preoperative gadoxetic acid-enhanced MRI and indocyanine green (ICG) test. Forty-five patients received open major hepatectomy. Liver function reserve and the future liver remnant were evaluated by computer-aided virtual hepatectomy. Global liver functional parameters included the T1 relaxation time reduction rate (T1_ratio) and functional liver volume (FV), whereas regional parameters included the rT1_pos, rT1_ratio, remnant FV (rFV), and remnant FV ratio (rFV_ratio) of the remnant liver. The functional parameters of the MRI and ICG were used to predict the short-term outcomes (liver failure and major complications) after major hepatectomy.

RESULTS

The T1_ratio and FV were correlated with the ICG test (rho = - 0.304 and - 0.449, p < 0.05). FV < 682.8 ml indicated preoperative ICG-_R15 ≥ 14% with 0.765 value of the area under the curve (AUC). No patient who underwent major resection with good liver functional reserve (ICG < 14%) and enough future remnant volume (> 30% standard LV) developed liver failure. Low rT1_ratio (< 66.5%) and high rT1_pos (> 217.5 ms) may predict major complications (AUC = 0.831 and 0.756, respectively; p < 0.05). The rT1_ratio was an independent risk factor for postoperative major complications (odds ratio [OR] = 0.845, 95% CI, 0.736-0.966; p < 0.05).

CONCLUSION

Preoperative gadoxetic acid-enhanced MRI with computer-aided virtual hepatectomy may facilitate optimal assessment of regional liver functional reserve to predict short-term outcomes after major hepatectomy for HCC.

KEY POINTS

• Preoperative gadoxetic acid-enhanced MRI with virtual hepatectomy and volumetric analysis can provide precise liver volume and regional functional assessment. • Quantitative regional liver function assessed by gadoxetic acid-enhanced MRI can predict the short-term outcomes after major hepatectomy in patients with HCC. • The regional liver function assessed by gadoxetic acid-enhanced MRI is an independent risk factor for postoperative major complications.

Leiomyosarcoma of the portal vein: a case report and review of the literature

Leiomyosarcomas (LMS) are rare malignant tumours of smooth muscle origin predominately affecting females in their sixth decade. Only 2% of LMS arise from blood vessels and most are in the inferior vena cava. We present the first reported case of LMS of the portal vein in a male patient. Multidetector CT showed a large mass in the main portal vein, which was initially misinterpreted as a pancreatic cancer. Careful examination of the multidetector CT images showed radiological features of an intraluminal mass, and a preoperative diagnosis of primary LMS of the main portal vein was made. The patient underwent curative surgery and made an uneventful recovery. Awareness of this entity and recognition of the salient CT features may facilitate radiologists in making the correct preoperative diagnosis.