Highlights

1. ARTIFICIAL INTELLIGENCE TRANSFORMS ADVANCED LIVER DISEASE CARE: FROM RISK PREDICTION TO PERSONALIZED DECISION-MAKING

Artificial intelligence (AI) is rapidly reshaping the management of cirrhosis and portal hypertension. This review summarizes how machine learning and other AI approaches are being applied to disease diagnosis, complication prediction, prognosis, and treatment-response assessment, while also addressing important challenges such as model generalizability, interpretability, workflow integration, and ethical considerations.¹ Looking ahead, the authors envision AI evolving from isolated prediction tools toward multimodal decision-support systems integrating clinical, imaging, laboratory, and omics data.¹

Why read it: This article provides a timely and clinically oriented overview of where AI may genuinely add value in advanced liver disease—and what still needs to be solved before AI can become part of routine patient care.

2. EVERY DECIBEL COUNTS: EARLY HEARING IMPROVEMENT PREDICTS COMPLETE RECOVERY AFTER SUDDEN HEARING LOSS

Can early hearing changes tell us who will ultimately recover from sudden sensorineural hearing loss? In this study of 240 patients, each 1-dB improvement within the first two weeks was associated with an approximately 1.5-percentage-point higher probability of complete hearing recovery.² Importantly, the prognostic value of early improvement varied according to the severity of hearing loss at presentation, with stronger predictive value in patients with less severe baseline impairment.

Why read it: The study turns a routinely available audiometric measurement into an intuitive and quantitative prognostic tool that may improve patient counseling, follow-up planning, and clinical decision-making.²

3. PROTON VS PHOTON THERAPY: WHAT REALLY DRIVES SKIN TOXICITY IN HEAD-AND-NECK RADIOTHERAPY?

Proton therapy offers superior dose distribution in many settings, but does it actually reduce the skin toxicity experienced by patients? This prospective cohort study compared 158 patients receiving intensity-modulated proton therapy or volumetric modulated arc therapy and incorporated both physician-graded radiation dermatitis and patient-reported skin-related quality of life.³ Skin toxicity worsened during treatment and improved afterward, but no significant advantage of proton therapy was observed. Instead, advanced nodal disease and higher radiation dose were the major predictors of more severe skin toxicity.³

Why read it: The findings challenge the assumption that a more advanced radiation modality automatically translates into better patient-reported skin outcomes and emphasize the importance of risk-adapted supportive care.

4. ENDOMETRIAL SCRATCHING DOES NOT IMPROVE PREGNANCY RATES IN FROZEN EMBRYO TRANSFER

Endometrial scratching has been widely adopted as an adjunctive fertility procedure despite ongoing uncertainty regarding its benefit. In this cohort of 822 women undergoing frozen embryo transfer, endometrial scratching was not associated with a significant improvement in clinical pregnancy rates.⁴ The findings did not support routine use in treatment-naïve women with a favorable prognosis, while possible benefit among women with repeated implantation failure remained uncertain because estimates were imprecise and statistically nonsignificant.

Why read it: This study addresses a common but controversial fertility intervention and provides clinically relevant evidence against unnecessary routine procedures, while clearly identifying the population in whom further randomized trials are still needed.⁴

5. WHO WILL EXPERIENCE PAIN DURING OFFICE HYSTEROSCOPY? TWO SIMPLE PREDICTORS IDENTIFY HIGH-RISK PATIENTS

Office hysteroscopy is convenient and minimally invasive, but procedural pain remains an important barrier for many women. Among 338 patients undergoing hysteroscopy without anesthesia, approximately 41% experienced moderate-to-severe pain.⁵ Severe dysmenorrhea and a cervical width of <4 mm emerged as particularly strong independent predictors, with adjusted odds ratios of 9.42 and 6.51, respectively.⁵

Why read it: Two simple clinical and anatomical features may help clinicians identify women at high risk for procedural pain before hysteroscopy, allowing more individualized counseling and targeted pain-management strategies.

CONFLICTS OF INTEREST

Dr. Jong-Ling Fuh, an editorial board member at the Journal of the Chinese Medical Association, had no role in the peer review process of or decision to publish this article.