Differential diagnosis, imaging and steroids

(Agent report, 2026-09-03. Verified by Kya at synthesis time — see biopsy-case doc.)

Q1 Differential & frequencies

• Razek & Castillo AJNR 2009 (PMC7051469): full CS differential; "MR imaging features of many of these lesions overlap and are often nonspecific". Lymphoma: "most have no specific MR imaging findings".
• Bozer & Pekcevik 2025 (PMC12591083): "No conventional CT or MRI feature reliably distinguishes chordomas from chondrosarcomas."
• Van Pee 2025 (PMC12700139): "when imaging findings are atypical, histopathology remains essential for definitive diagnosis."
• Keane 1996 (151 CS syndrome): tumours 30%, trauma 24%, self-limited inflammation 23% (mean age 35).
• Fernandez 2007 (126): tumours 63%; pain at onset OR 12 for Tolosa-Hunt; male sex/V2/no pain favour tumour.
• Dellaretti 2017 SR of 75 percutaneous CS biopsies: 58 neoplastic (26 meningioma), 6 non-neoplastic.
• Nagaraja 2022 (38 resected isolated CS lesions): hemangioma 50%, aspergilloma 13%, meningioma 13%, schwannoma 8%, mets 8%, chondrosarcoma 5%, chordoma 3%. 29.4% not accurately diagnosed on imaging.
• Zhang 2023 (11 endoscopic biopsies of undifferentiated CS lesions): 10/11 diagnosed; 6 got non-RT treatment.
• Chordoma: young age (<26) & female -> cranial presentation (McMaster SEER). Wang 2017 n=238 mean age 38.

Q2 Treatment-changing alternatives

• Potter Radiographics 2024: "Essential requirements for PBT following gross total or maximal safe resection are tissue diagnosis..."
• Lymphoma: EHA-ESMO 2024 (PMC11148853) chemo-immunotherapy; Skotting 2024 skull-base surgery index complication rate 51.5%.
• Plasmacytoma: definitive RT (Tsang ILROG 2018). NPC: chemoRT (ESMO 2021). Prolactinoma: dopamine agonist. Tolosa-Hunt/IgG4: steroids. Infection incl aspergilloma: antimicrobials.
• Messerer 2012: biopsy "would enable patients to obtain the most appropriate therapy, such as resective surgery, corticosteroids, chemotherapy, radiotherapy, or radiosurgery."
• Chondrosarcoma: Almefty 2007 better outcome, RT may be unnecessary for low grade — least harmful misdiagnosis.
• PDC: Duan 2022 Ki-67 20-50%, OS 2-17 months; overwhelmingly paediatric (median 7).

Q3 Imaging accuracy

• Yeom AJNR 2013: ADC chondrosarcoma 2051, classic chordoma 1474, PDC 875. Welzel 2018 ADC cut-off 1585 sens 94%/spec 99% — but only among confirmed chordoma/chondrosarcoma.
• Yamazawa 2022: neurosurgeons 40-80% (median 60%) at chordoma vs chondrosarcoma on conventional MRI; "radiographical twins".
• Imaging wrong in ~15-30% of CS surgical series. Low ADC doesn't exclude lymphoma/PDC.

Q4 Steroids before biopsy

• EHA-ESMO 2024: "Corticosteroid therapy before tissue biopsy should be avoided whenever clinically possible [IV, D]"; "In case of clinical deterioration, urgent biopsy should be carried out before the start of corticosteroids [IV, A]".
• BSH 2019 (snippet): non-diagnostic 33% after <1 wk steroids, 57% after longer.
• Tosefsky 2024 meta-analysis (19 studies, 1226 pts): RR non-diagnostic 2.1 (1.1-4.1); stereotactic subgroup RR 3.0.
• Scheichel 2021 pooled: 10.3% vs 3.4% inconclusive, OR 3.3.
• Steroid response NOT diagnostic: Dutta 2021; Ruff 2018 (CS chordoma clinically steroid-responsive); Demirkaya 2010 (lymphoma called Tolosa-Hunt).

Q5 Growth rate

• Wang 2020 (21 untreated SB chordomas): specific growth rate 8%±9%/month, 0.8±0.7 cm3/month; "slow-growing". Pallini 2003: ~20% recur within 1 yr.
• Inference: ~8 cm3 in 8 months is at the extreme end for conventional chordoma; favours missed-in-January, PDC, lymphoma, mets, or inflammatory.

Q6 Jan MRI miss plausibility

• Sakai 2023 (PMC10066091): non-FS post-contrast T1 lesion signal can equal marrow fat; ~20% of "brain MRI" mets are non-brain. Bello Radiographics 2019: skull base easily overlooked on first/last images. Clivus "under-recognized" (EPOS 2025). Skull-base protocol <=3mm slices, fat-sat, DWI — not routine brain protocol.
• Action: neuroradiologist re-read of Jan study for clival marrow / CS asymmetry; confirm contrast, fat-sat, coronal thin slices, FOV.

Synthesis

Tissue required for PBT anyway (Potter 2024; Stacchiotti 2015 "preoperative histological diagnosis is recommended in principle", skull-base exception presumes histology at resection). Atypical features shift probability to treatment-changing diagnoses. Imaging can't close the gap. Asymmetric harm. Sequence steroids after tissue. Counterpoints: consensus allows skipping dedicated pre-op biopsy when access hard/seeding risk; biopsy not risk-free (Messerer 14% unproductive; Dellaretti 3 permanent deficits/75). Framing: biopsy is not an alternative to treatment but the first step of it.

Addendum 27 Sep 2026: plasmacytoma ADC (verified PubMed abstracts)

• Hwang 2021, Diagn Interv Imaging, PMID 34127433: bone plasmacytoma mean ADC 760.1 +/- 196.9 um2/s vs bone metastases 1214.2 +/- 382.6; ADC <= 908.3 gave sensitivity 88% and specificity 78% for plasmacytoma; AUC 0.913.
• Duan 2022, J Comput Assist Tomogr, PMID 35297583: sinonasal extramedullary plasmacytoma, RESOLVE-DWI, mean ADC 0.55 x10-3 mm2/s (n=9).
• Miladinovic 2025, Skeletal Radiol, PMID 38967687: skull base, spine and sacrum primary bone neoplasms; chondrosarcoma had the highest ADC, significantly above chordoma. Not read in full.
• The patient (my measurement, about 0.85) falls inside both the plasmacytoma and the PD chordoma (Yeom 875 +/- 100) ranges.