At a glance
From Parkinson's first surgical route to Knosp grading, radiosurgery and the endoscopic correction, ten cavernous sinus papers reshaped how clinicians diagnose and treat the region.
The cavernous sinus is a small thing to have caused so much trouble. Roughly two centimetres across, it is a venous channel on either side of the sphenoid bone, threaded by the internal carotid artery and by the nerves that move the eye. For most of the twentieth century the standard account of it was an account of hazards: the space was described at length so that surgeons would know what not to touch.
What changed was a run of papers, concentrated between 1965 and 2015, that turned a region into a set of named compartments, grades and target volumes. The list below is a working one, and it selects for influence on what a clinician does in a clinic rather than for how often a paper is cited. A study earns its place here if it changed a decision: whether to operate, how to grade a scan, when to irradiate, what to tell a patient at the end of a consultation.
A venous channel that behaves like a crossroads
The modern description divides the space into triangles, and the triangle names are the vocabulary of every operative note written about the region. Parkinson's triangle, the infratrochlear space, is bounded by the trochlear nerve, the ophthalmic division of the trigeminal nerve and the tentorial edge; it holds the cavernous carotid and the abducens nerve, and opening it allows the sixth nerve to be followed from Dorello's canal to the superior orbital fissure. The oculomotor, or Hakuba's, triangle reaches the horizontal segment of the carotid and tumours in the medial compartment. The supratrochlear triangle exposes the meningohypophyseal trunk.
Size explains the rest of the clinical picture. Watanabe's measurements of Parkinson's space put its three sides at roughly 13, 16 and 6 millimetres. A mass of a few millimetres, or a tear in the arterial wall, therefore produces symptoms out of proportion to its bulk, and the signature is the one Tolosa and Hunt described: painful ophthalmoplegia with cranial nerve palsies.
1954 to 1983: a diagnosis, a map, and the first instruments
The first obstacle was not technical. It was that nobody could say with confidence what was wrong.
Tolosa, then Hunt: an inflammation that answers to steroids
Eduardo Tolosa's 1954 report in the Journal of Neurology, Neurosurgery and Psychiatry described a granulomatous reaction around the carotid siphon in a patient whose presentation had mimicked an infraclinoid aneurysm. Seven years later, Hunt and colleagues argued in Neurology that the underlying process was an indolent, non-specific inflammation of the cavernous sinus itself, and that it improved with corticosteroids. The eponym that followed gave physicians a diagnosis they could treat without an operation, and it is still made the same way — by exclusion, on the strength of a painful ophthalmoplegia that settles when steroids are given.
Parkinson: the cavernous carotid becomes reachable
Until 1965, as later reviews of the field record, the neurosurgical literature contained almost nothing on direct operative attack on lesions of the cavernous sinus, a gap produced by what uncontrolled bleeding and cranial nerve injury meant before the microscope. Donald Parkinson's anatomical study and case report in the Journal of Neurosurgery changed the terms of that argument by demonstrating a route in. The triangle that carries his name is still the standard door.
Serbinenko: balloons, and the shift inside the artery
Fedor Serbinenko, working in Russia, had developed a flow-directed balloon catheter in 1962 and published his experience in the Russian literature in 1971; the English-language account appeared in the Journal of Neurosurgery in 1974. The balloons ran from half a millimetre to 1.5 millimetres, steered under fluoroscopy, and were used to occlude vessels, to test collateral flow through the circle of Willis and, in a remark almost buried in the paper, to occlude aneurysms. The mortality figures were not hidden: two deaths in 304 temporary occlusions, and two in 162 permanent occlusions, both after catheterisation of the basilar artery. A retrospective commentary in the American Journal of Neuroradiology notes that the paper has had great implications for the advancement of neurointerventional technique.
Harris and Rhoton: the compartments get named
The microsurgical study Harris and Rhoton published in the Journal of Neurosurgery in 1976 belongs to the laboratory wave that the field's own historiography credits with making medial exploration of the cavernous sinus feasible. Its contribution was descriptive and it has lasted: the venous plexus as a set of compartments arranged around the carotid, the dural walls, the nerves in their beds. Later grading schemes assume that description rather than argue with it.
Dolenc: the combined approach
Vinko Dolenc's 1983 paper on direct microsurgical repair of intracavernous vascular lesions is the one that made the space routinely operable rather than occasionally so. The route he developed — pterional craniotomy, division of the meningo-orbital band, extradural removal of the anterior clinoid process, drilling of the optic strut, unroofing of the optic canal, then a split of the dura at the superior orbital fissure to reach the lateral wall from between its layers — became the standard transcranial method, and it remains the one against which endoscopic alternatives are still measured.
1985 to 2015: classification, radiation, and a correction
By the mid-1980s the technical question had narrowed. The open questions were which lesions to treat, and with what.
Barrow: four types of fistula
Carotid-cavernous fistulas had been discussed as a single disease. Daniel Barrow and colleagues, reporting 14 patients with spontaneous fistulas in the Journal of Neurosurgery in 1985, split them by anatomy and angiographic supply, and the split has governed treatment ever since.
- Type A: a direct, high-flow shunt between the internal carotid artery and the cavernous sinus.
- Type B: dural shunts fed by meningeal branches of the internal carotid artery.
- Type C: dural shunts fed by meningeal branches of the external carotid artery.
- Type D: dural shunts fed by meningeal branches of both.
The paper set out the clinical manifestations, the angiographic evaluation, the indications for therapy and the options for each type. That structure is why a clinician faced with a red, proptotic eye can say within minutes which vessel is feeding the problem.
Sekhar and Møller: tumours, and the case for reconstruction
Neoplasms were harder than fistulas, because the lesion sits in the wall rather than in the flow. Laligam Sekhar's Pittsburgh group, reporting in the Journal of Neurosurgery in 1986 on the operative management of tumours involving the cavernous sinus, set out an approach to lesions that had previously been declined. A companion anatomical study published the following year was explicit about the pairing at the heart of the method: operative approaches with related vascular and neural reconstruction. Surgery could be offered because the artery and the nerves could, in some measure, be rebuilt.
Knosp: five grades drawn on the carotid
Engelbert Knosp and colleagues, then in Mainz, made lateral extension of pituitary adenomas measurable. Their 1993 classification in Neurosurgery used the internal carotid artery as its landmark on coronal MRI: a medial tangent, an intercarotid line drawn between the cross-sectional centres of the intra- and supracavernous carotid, and a lateral tangent separate grades 0 to 4, with grade 4 representing complete encasement of the intracavernous carotid. The validation was surgical rather than radiological. All grade 3 and grade 4 tumours proved invasive, as did all but one grade 2 tumour, while grades 0 and 1 did not. The same paper reported a statistically higher Ki-67 proliferation index in invasive adenomas. The scheme stuck because of where it placed the boundary: the zone between the intercarotid line and the lateral tangent is precisely where a surgeon must decide how much to chase. The original classification remains the reference point for every later revision.
Lee and colleagues: radiosurgery as the primary option
Two decades of argument about whether cavernous sinus meningiomas should be resected ended, in practice, with a report from Pittsburgh covering 14 years of experience. Niranjan, Kondziolka, Flickinger, Lunsford and colleagues reviewed 176 patients with meningiomas centred in the cavernous sinus; 159 were analysable, and 83 had radiosurgery as their only treatment. At a median margin dose of 13 Gy, tumour control for typical meningiomas was 93.1 per cent at five years and the same at ten. For the primary-radiosurgery group it was 96.9 per cent at five years. Neurological status improved in 29 per cent of patients and stayed stable in 62 per cent. Adverse radiation effects followed 11 procedures, or 6.7 per cent. The authors concluded that radiosurgery was the preferred strategy for tumours up to about three centimetres in diameter or 15 cubic centimetres in volume.
Micko: the endoscope audits the grade
The correction came from inside Knosp's own group. Alexander Micko, with Wöhrer, Wolfsberger and Knosp, took 137 pituitary macroadenomas graded on MRI and looked at them through an endoscope in the operating room. Overall, 16 per cent were invading. Grade 1 adenomas were invasive in 1.5 per cent of cases, grade 2 in 9.9 per cent, grade 3 in 37.9 per cent. The consequential finding was the split inside grade 3: tumours involving the superior compartment invaded in 26.5 per cent of cases, with gross-total resection or endocrine remission achieved in 85 and 67 per cent, while tumours involving the inferior compartment invaded in 70.6 per cent and endocrine remission was zero. Every grade 4 adenoma was invasive, with no gross-total resection and no remission.
The paper also did something less comfortable. Comparing endoscopic with microscopic assessment, it found that grade 2 tumours had been recorded as invasive in 88 per cent of cases under the microscope and in 9.9 per cent under the endoscope; for grade 3 the figures were 86 and 37.9 per cent. The published numbers had overstated invasion, which means they had also overstated inoperability. The revised scheme added grades 3A and 3B on the strength of that discrepancy.
The objection that the list is arbitrary
The strongest case against this kind of ranking is that it confuses influence with accessibility. Much of the evidence behind these papers is retrospective and single-centre. Serbinenko's first report was in Russian, and the English version arrived three years later. Surgical technique papers date quickly once a new corridor opens, and citation counts reward any paper that supplies a classification others are obliged to cite, whether or not the classification survives contact with a second operating theatre.
It does not follow, however, that the list is decorative. The Micko paper is the answer to the objection, because it shows the canon correcting itself: a classification from 1993 was tested by the very group that wrote it and found to have overstated invasion by a wide margin. External work pointed the same way. A 2019 series of 254 patients published in Pituitary found cavernous sinus invasion in 18.4 per cent overall, with gross-total resection in 55.8 per cent of grade 3A tumours against 30.0 per cent of grade 3B tumours — the same split, found independently, and a reminder that a grade is a hypothesis until someone else's patients confirm it.
Where the argument sits now
Practice has converged, unevenly, on treating the region in layers. The European Association of Neurosurgical Societies' skull base section published a consensus statement on the management of cavernous sinus meningiomas in 2022, and the fact that a consensus statement was thought necessary is itself evidence of how thin the randomised evidence remains. The exception that keeps surgeons in the room is the optic apparatus. A tumour pressing on the optic nerves cannot safely be irradiated to the doses the tumour needs, so debulking to create space, or to relieve severe retro-orbital pain, remains an indication even where radiosurgery leads.
Fistulas have moved almost entirely to the endovascular suite, pituitary adenomas are graded preoperatively and resected through the nose where possible, and the open transcranial approach is now a reserve. None of that would have been legible to a clinician reading the literature in 1970.
The precedent a grade sets
A classification is not a neutral description of nature. Once Knosp's grades were printed, they entered the consent conversation: a patient with an inferior-compartment adenoma is told, in effect, that remission is unlikely, because one series of 137 operations showed it. That is a heavy load for a single-centre study to carry for three decades, and it is a burden the 2015 revision made lighter rather than heavier. The 2025 history of skull base surgery published in Acta Neurochirurgica traces how a region once treated as untouchable became mapped, entered and reconstructed. It reads as a story of instruments. It is at least as much a story of nomenclature outrunning evidence.
What precedent does that set for the next revision? The likeliest is that image-based prediction — a scan that forecasts invasion before an endoscope confirms it — will replace the tangent lines drawn on coronal MRI. The test of that work will not be its citation count. It will be whether the numbers hold when a different theatre, in a different country, checks them, and whether a clinician is willing to say them out loud to a patient.
Photo by Abdulai Sayni on Unsplash
- cavernous sinus
- neurosurgery
- Knosp grading
- carotid-cavernous fistula
- stereotactic radiosurgery

