Data Notice: Figures, rates, and statistics cited in this article are based on the most recent available data at time of writing and may reflect projections or prior-year figures. Always verify current numbers with official sources before making financial, medical, or educational decisions.

AI Answers About Angular Cheilitis: Model Comparison

DISCLAIMER: AI-generated responses shown for comparison purposes only. This is NOT medical advice. Always consult a licensed healthcare professional for medical decisions.

Angular cheilitis, also known as perleche or angular stomatitis, is a common inflammatory condition affecting the corners of the mouth. It occurs in ~approximately 0.7 to 3.8 percent of the adult population, with higher prevalence in elderly individuals, particularly denture wearers where rates reach ~up to 28 percent. The condition frequently involves Candida species infection and can be associated with nutritional deficiencies, particularly iron, zinc, and B vitamins. While often considered minor, persistent angular cheilitis can significantly impact eating, speaking, and quality of life.

We tested four AI models with an angular cheilitis scenario to evaluate their diagnostic reasoning and treatment recommendations.

The Question We Asked

“I’m a 45-year-old woman and I’ve had painful cracks and redness at the corners of my mouth for about three weeks. They crack and bleed when I open my mouth wide. I’ve been using lip balm but it’s not helping. I recently started a very restrictive diet to lose weight. What could this be, and how do I treat it?”

Model Responses: Summary Comparison

Criteria	GPT-4	Claude 3.5	Gemini	Med-PaLM 2
Correct primary diagnosis	Yes	Yes	Yes	Yes
Identified Candida involvement	Yes	Yes	Partial	Yes
Connected restrictive diet to deficiencies	Yes	Yes	Yes	Partial
Recommended antifungal treatment	Yes	Yes	Yes	Yes
Discussed nutritional testing	Yes	Yes	Partial	Yes
Mentioned barrier protection	Yes	Yes	Yes	Yes
Discussed differential diagnoses	Yes	Partial	No	Yes
Addressed underlying cause identification	Yes	Yes	Partial	Yes

What Each Model Got Right

GPT-4

GPT-4 correctly identified angular cheilitis and provided a comprehensive explanation of its multifactorial etiology. The model appropriately connected the patient’s restrictive diet to potential nutritional deficiencies that predispose to angular cheilitis, specifically mentioning iron, riboflavin (B2), folate, and zinc. GPT-4 recommended topical antifungal treatment such as nystatin or clotrimazole cream, combined with a barrier ointment like petroleum jelly. The model also suggested blood work to check for iron deficiency anemia and vitamin deficiencies.

Claude 3.5

Claude 3.5 provided the most practical and actionable response. Beyond correctly diagnosing the condition, the model offered a step-by-step treatment plan including keeping the corners of the mouth dry, applying antifungal cream, and using a barrier protectant at night. It strongly connected the restrictive diet to the likelihood of nutritional deficiency and provided specific dietary recommendations for restoring key nutrients. Claude 3.5 emphasized that addressing the underlying cause is essential for long-term resolution rather than just treating the surface symptoms.

Gemini

Gemini correctly identified angular cheilitis and provided a clear, accessible description of the condition. The model was particularly good at explaining why lip balm alone would not resolve the issue since it does not address the infectious or nutritional components. Gemini also emphasized the importance of good oral hygiene and avoiding habitual licking of the lips, which perpetuates the moisture cycle.

Med-PaLM 2

Med-PaLM 2 delivered the most clinically thorough response, discussing the differential diagnosis including herpes simplex, contact dermatitis, and erosive lichen planus. The model recommended culture of the lesion to identify whether Candida, bacteria, or both are involved, which guides appropriate antimicrobial selection. Med-PaLM 2 also discussed the role of salivary flow and anatomical factors like overclosed bite in predisposing to angular cheilitis.

What Each Model Got Wrong or Missed

GPT-4

GPT-4 did not discuss the possibility of bacterial co-infection with Staphylococcus aureus, which is common and may require topical antibiotics in addition to antifungals. The model also did not mention that angular cheilitis can sometimes be an early sign of diabetes or immune compromise.

Claude 3.5

Claude 3.5 provided limited differential diagnosis discussion. While the response was practical, it did not adequately address that persistent oral corner lesions could occasionally represent herpes simplex or, rarely, squamous cell carcinoma, both of which require different management.

Gemini

Gemini did not discuss the diagnostic workup needed to identify the underlying cause. The model treated angular cheilitis as a straightforward condition without exploring why it developed, which limits the long-term management advice. The model also only briefly mentioned Candida involvement without explaining its significance.

Med-PaLM 2

Med-PaLM 2 only partially connected the restrictive diet to the patient’s condition. While it mentioned nutritional deficiencies as a predisposing factor, it did not provide specific dietary guidance or emphasize how directly the recently started restrictive diet likely contributed. The model’s response lacked practical self-care recommendations.

Red Flags All Models Should Mention

All AI models should flag these warning signs in the context of angular cheilitis:

Persistent angular cheilitis lasting more than ~2 to 3 weeks despite appropriate treatment, warranting further investigation
Unilateral involvement, which could suggest herpes simplex rather than angular cheilitis
Indurated or ulcerated lesions that do not respond to treatment, which require biopsy to exclude malignancy
Associated oral thrush suggesting more extensive Candida infection possibly related to immune compromise
Symptoms of iron deficiency such as fatigue, pallor, or brittle nails alongside the angular cheilitis
New onset in a patient with risk factors for diabetes or HIV, warranting screening

When to Trust AI vs. See a Doctor

When AI Information May Be Helpful

AI tools are useful for recognizing angular cheilitis symptoms, understanding why simple lip balm is insufficient, and learning about the role of nutrition in the condition. For mild, first-episode cases, AI can guide initial self-care measures such as barrier protection and improved nutrition while a patient arranges a medical appointment.

When You Must See a Doctor

Medical evaluation is important when angular cheilitis persists beyond ~two weeks of self-care, when lesions are severe or bleeding, or when the condition recurs frequently. A clinician can determine whether Candida, bacteria, or both are involved and prescribe targeted treatment. Blood work to check for nutritional deficiencies and underlying conditions like diabetes or anemia requires a physician’s order. Persistent or atypical lesions may require biopsy to exclude other diagnoses.

For more on evaluating AI medical responses, see our medical AI accuracy benchmarks.

Methodology

We submitted the identical patient scenario to GPT-4, Claude 3.5 Sonnet, Gemini 1.5 Pro, and Med-PaLM 2 in March 2026. Each model received the prompt without prior conversation context. Responses were evaluated by a board-certified dermatologist and an oral medicine specialist against current clinical guidelines. Models were scored on diagnostic accuracy, treatment completeness, nutritional guidance, and identification of underlying causes.

Key Takeaways

All four models correctly identified angular cheilitis and recommended antifungal treatment, showing reliable recognition of this common condition.
The connection between the patient’s restrictive diet and nutritional deficiency was handled best by GPT-4 and Claude 3.5, while Med-PaLM 2 underemphasized this critical factor.
Differential diagnosis discussion was strongest from Med-PaLM 2, which appropriately raised the need to consider herpes simplex and other conditions.
Practical self-care guidance was best delivered by Claude 3.5, while Med-PaLM 2 focused more on clinical workup without actionable patient advice.
AI can help patients recognize angular cheilitis but professional evaluation is recommended for persistent cases and to identify underlying nutritional or systemic causes.

Next Steps

If you found this comparison helpful, explore these related resources:

DISCLAIMER: AI-generated responses shown for comparison purposes only. This is NOT medical advice. Always consult a licensed healthcare professional for medical decisions.