Gallbladder carcinoma in the era of artificial intelligence: Early diagnosis for better treatment

Abstract

Gallbladder carcinoma (GBC) is the most common malignant tumor of biliary tract, with poor prognosis due to its aggressive nature and limited therapeutic options. Early detection of GBC is a major challenge, with most GBCs being detected accidentally during cholecystectomy procedures for gallbladder stones. This letter comments on the recent article by Deqing et al in the World Journal of Gastrointestinal Oncology, which summarized the various current methods used in early diagnosis of GBC, including endoscopic ultrasound (EUS) examination of the gallbladder for high-risk GBC patients, and the use of EUS-guided elastography, contrast-enhanced EUS, trans-papillary biopsy, natural orifice transluminal endoscopic surgery, magnifying endoscopy, choledochoscopy, and confocal laser endomicroscopy when necessary for early diagnosis of GBC. However, there is a need for novel methods for early GBC diagnosis, such as the use of artificial intelligence and non-coding RNA biomarkers for improved screening protocols. Additionally, the use of in vitro and animal models may provide critical insights for advancing early detection and treatment strategies of this aggressive tumor.

Key Words: Gallbladder carcinoma; Endoscopic ultrasound; Biopsy; Elastography; Choledochoscopy; Artificial intelligence; Non-coding RNAs; Screening; Animal models; In vitro studies

Core Tip: Gallbladder carcinoma is the most common malignant tumor of biliary tract, with poor prognosis due to its aggressive nature and limited therapeutic options. Current methods available for early screening include endoscopic ultrasound (EUS) examination, EUS-guided elastography, contrast-enhanced EUS, trans-papillary biopsy, natural orifice transluminal endoscopic surgery, magnifying endoscopy, choledochoscopy, and confocal laser endomicroscopy. Despite these established modalities, there is a need for innovative diagnostic methods, particularly the use of artificial intelligence and non-coding RNA biomarkers, to improve screening protocols and facilitate earlier disease detection.

TO THE EDITOR

Gallbladder carcinoma (GBC) is the most common biliary malignant tumor with an aggressive nature and poor prognosis[1,2]. The editorial reported by Deqing et al[1]: “Endoscopic diagnosis and management of Gallbladder carcinoma in a minimally invasive era: New needs, new models”, highlighted various minimally invasive diagnostic and therapeutic procedures in GBC. This editorial written by Deqing et al[1] was based on the work published by Pavlidis et al[2]: “New Trends in the Diagnosis and Management of Gallbladder Carcinoma”. GBC is diagnosed late and has a poor prognosis with a 5-year survival rate of 13%[3].

Imaging, endoscopic and non-endoscopic methods for early detection of GBC

Deqing et al[1] emphasized the early detection of GBC and recommended incorporating endoscopic ultrasound (EUS) examination of the gallbladder inner wall as a quality control indicator for high-risk GBC patients. They also advocated using EUS-guided elastography, contrast-enhanced EUS, trans-papillary biopsy, natural orifice transluminal endoscopic surgery, magnifying endoscopy, choledochoscopy, and confocal laser endomicroscopy, when necessary for early diagnosis of GBC[1].

Other new emerging modalities used for differentiating benign and malignant lesions in the liver and breast such as real-time elastography using acoustic radiation force impulse (ARFI) could play an important role in the diagnosis of GBC[4,5]. Routine use of ARFI while doing ultrasonography combined with multidimensional computed tomography (CT) and contrast-enhanced endoscopic ultrasonography can help diagnose and stage early diffuse wall thickening type GBC[6].

The use of dual-time-point 18F-fluoro-2-deoxy-2-D-glucose [(18)F-FDG] positron emission tomography (PET) has also shown encouraging results in predicting gall bladder malignancy in gall bladder polyps. Delayed (18)F-FDG PET is more helpful than early (18)F-FDG PET for evaluating malignant lesions because of increased lesion uptake and increased lesion-to-background contrast, and when combined with the retention index, sensitivity is increased to 100% and specificity to 80%[2,7].

Early diagnosis of GBC using artificial intelligence and non-coding RNAs

Recent investigations have examined the potential of artificial intelligence (AI) in enhancing the detection and early diagnosis of GBC through CT imaging analysis. Deep learning AI algorithms have demonstrated particularly promising results, with improved efficacy and accuracy at early diagnosis of GBC, achieving accuracy rates of 98.35% using deep neural networks and MobileNet architectures. Conventional neural network approaches have shown moderate efficacy, with an area under the curve of 0.81. Despite their considerable promise, several barriers to clinical implementation persist, including regulatory compliance requirements, ethical considerations, as well as the need for validation and integration within existing clinal workflows. Despite these challenges, the integration of AI in CT scan analyses presents a promising method for the early diagnosis of GBC[8].

The early identification of GBC is a major clinical challenge, with most cases being discovered accidentally during cholecystectomy procedures performed for gallbladder stones. At present, there is a lack of effective population- level screening tests for GBC. Traditional tumor markers, including carcinoembryonic antigen and carbohydrate antigen 19-9 demonstrate limited diagnostic ability, particularly in early-stage GBC[9]. As a result, non-coding RNAs have emerged as potential novel biomarkers for early diagnosis and treatment of GBC, given their role in transcriptional regulation of target genes associated with solid tumor development[10,11]. However, there is a need for further research using both in vitro systems and experimental animal models to identify and validate new GBC-specific biomarkers[12].

The editorial by Deqing et al[1] presents limitations in its analysis, notably omitting crucial economic considerations related to novel GBC diagnostic technologies. A comprehensive evaluation should address the healthcare workforce and time costs, as well as medical insurance reimbursement implications, factors which are essential to assess feasibility of clinical implementation and integration into current systems to enhance the practicality and clinical relevance of their recommendations[13-15].

CONCLUSION

Current diagnosis methods for GBC cover a spectrum of techniques, such as EUS examination, EUS-guided elastography, contrast-enhanced EUS, trans-papillary biopsy, natural orifice transluminal endoscopic surgery, magnifying endoscopy, choledochoscopy, and confocal laser endomicroscopy. Despite the variety of diagnostic techniques available, early detection of GBC remains a major challenge, with most GBCs being detected accidentally during cholecystectomy procedures for gallbladder stones. This challenge has prompted the study of new modalities for early detection of GBC, particularly into AI and non-coding RNA biomarkers to improve screening protocols for GBC. Additionally, further research using in vitro systems and animal models of GBC is essential to develop novel diagnostic approaches and detect new biomarkers to assist with early diagnosis and better treatment of this aggressive tumor.