夜雨聆风A 78 years old caucasian male presented with new onset left retroorbital pain for the past two months. The pain has been dull and constant on a daily basis, and is worse in the morning. He developed Photosensitivity with headaches and was treated for migraines. Topamax did not help. He had an MRI of the brain with and without contrast which is negative. He went to the ER and was admitted. Ophthalmology and neurology were also consulted. No conclusive diagnosis was made.What is a possible diagnosis?
He is very intelligent and knowledgeable. Since he mentioned AI, I checked AI after I saw him this morning at clinic. I was surprised that AI did not give me a right answer among all its differential diagnosis. Here is the AI finding:
In a 78-year-old man with daily unilateral retro-orbital pain worse in the morning and photophobia, the failure of topiramate (Topamax, Eprontia) and negative brain MRI should prompt urgent consideration of giant cell arteritis (GCA) as the leading diagnostic concern [1] [2]. New-onset headache after age 50 carries a substantially elevated risk of secondary causes, with GCA specifically requiring prompt evaluation to prevent irreversible vision loss [3] [4]. The morning predominance of pain and photophobia are compatible with GCA, though jaw claudication, scalp tenderness, or visual symptoms may be absent in up to one-third of cases [5] [6].
Differential diagnosis
1. Giant cell arteritis (most likely) — The 2022 ACR/EULAR classification criteria require age ≥50 years plus a score ≥6 from: morning stiffness in shoulders/neck (+2), sudden visual loss (+3), jaw/tongue claudication (+2), new temporal headache (+2), scalp tenderness (+2), abnormal temporal artery exam (+2), ESR ≥50 mm/h or CRP ≥10 mg/L (+3), positive temporal artery biopsy or halo sign on ultrasound (+5), bilateral axillary involvement (+2), or FDG-PET aortic activity (+2) [1]. Notably, up to 20% of biopsy-proven GCA cases have normal inflammatory markers at presentation [6].
2. Cluster headache — Strictly unilateral severe orbital/supraorbital/temporal pain lasting 15–180 minutes, occurring 1–8 times daily, with ipsilateral autonomic features (conjunctival injection, lacrimation, nasal congestion, ptosis/miosis) and restlessness [7] [8]. The described daily pattern without remission periods would suggest chronic cluster headache, though the absence of autonomic features and the patient's age make this less probable.
3. Tolosa-Hunt syndrome — Unilateral orbital or periorbital pain with paresis of cranial nerves III, IV, and/or VI due to granulomatous inflammation in the cavernous sinus, superior orbital fissure, or orbit [7]. MRI with gadolinium typically shows enhancing inflammatory tissue; cases with initially normal MRI have been reported [9].
4. Painful optic neuritis — Retro-orbital pain with eye movement and visual impairment; gadolinium-enhanced MRI shows optic nerve enhancement in ≈90% of cases [7]. The absence of reported visual changes makes this less likely.
5. Hemicrania continua — Continuous unilateral pain with exacerbations and ipsilateral autonomic features, absolutely responsive to indomethacin [10]. The lack of autonomic features and the patient's age reduce probability.
6. Trochleitis — Periocular pain emanating from the trochlear area, exacerbated on examination and looking in supraduction; corticosteroid injection provides rapid relief [11] [12].
7. Medication overuse headache — Daily headache in the setting of frequent analgesic use; requires ≥15 days per month of simple analgesics or ≥10 days per month of triptans/opioids for 3 months [13].
AI is right on the most important thing which is to rule out Temporal arteritis(GCA).This patient has no visual symptoms and was seen by both neurology and ophthalmology with a negative CRP/ESR.Of course both neurology and ophthalmology would not miss GCA or they all deserve to lose their licenses. LOL
For reassurance, I pressed his left temporal area and he clearly told me no pain, neither the other side. I also palpated all the scalp area and neck, shoulder area. No tenderness at all. Then I put both my hands on his bilateral temporal area and pushed down harder. Still no pain. I let him open the mouth, he did and then suddenly screamed. Severe pain comes from the left side of jaw.
I did not say anything and he did not say anything, neither any questions from him.It is obvious:
He has left TMJ.TMJ caused his retroorbital pain.
I guess he will AI the relationship between retroorbital pain and TMJ after he goes home. I wanted him to do it.
The real question is : why did AI not even mention it in the differential diagnosis?
AI can not replace doctors, never.
