This study was approved by the Ethics Committee of the Tangdu Hospital of Air Force Medical University (AFMU; Xi’an, China) and the Mianyang Central Hospital (MCH; Mianyang, China). We reviewed 17 patients who underwent surgical treatment at the Tangdu Hospital of AFMU and at MCH between January 2010 and November 2021.

Preoperative computerized tomography (CT), computerized-tomography angiography (CTA), and magnetic resonance imaging (MRI) were used to evaluate the sizes of ICDs and their relationships with surrounding nerves and blood vessels. Tumor volumes were calculated via MRI using the following formula:

Imaging examination was reviewed by two experienced personnel, with a third person to evaluate if there was any disagreement between the two personnel.

All 17 patients had primary tumors and underwent one operation. Surgical approach was based on tumor location on MRI. The pterion approach was used to treat the lesions located in the sellar region and extending into the anterior and middle cranial fossa, while the orbitozygomatic approach was used to remove the tumors located in the parasellar region and extending into the middle and posterior cranial fossa and the upper slope. The subfrontal approach was mainly aimed at resecting anterior skull base tumors. The EETA was mainly applied for sellar tumors. For large tumors extending into suprasellar, parasellar and petroclival regions, the EEETA was required. In each patient undergoing EETA or EEETA, a nasal-septum mucosal flap was created to cover the dural defect. We used neuronavigation and Doppler to locate the internal carotid artery (ICA) and its branches during the operation. Occasionally, tumors were removed with the aid of a plasma knife.

The extent of tumor resection was assessed on the basis of postoperative MRI and CT and was defined as follows: gross total resection (GTR), no tumor residue; subtotal resection (STR),>90% of tumor removed; and partial resection (PR), 50-90% of tumor removed.; and partial resection (PR), <90% of tumor removed.

Our literature search was based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses(PRISMA) (6). We systematically searched reports in the English language from PubMed/MEDLINE and EMBASE databases since 1990 using the Boolean operators “OR” “AND” in combination OR with the following keywords: “chondromas,” “chondroma,” “intracranial,” “skull base,” “brain,” “sella,” and “sellar.” All identified articles were screened in two stages. Firstly, titles and abstracts were screened for relevant studies. Secondly, the full text was downloaded and assessed for eligibility. This process was carried out independently by three assessors. Any disagreements were resolved by consensus.

The clinical data of 17 patients for the ICDs are summarized in Table 1 . Thirteen patients were female and four were male, ranging from 15 to 61 years of age (average, 33.7 years). The duration of clinical symptoms in patients ranged from 0.3 to 36 months (average, 8.6 months). Patients presented with varying clinical symptoms, including headache (5/17), dizziness (1/17), hypopsia (6/17), diplopia (3/17), facial numbness (2/17), ptosis (1/17), limb numbness (1/17), and sexual dysfunction (1/17; Table 1 ). Headache and hypopsia were the most common symptoms in the study. The possible cause of headache in five cases was elevated intracranial pressure (ICP) due to a large space-occupying lesion or the tumor irritating the dura mater. Hypopsia, diplopia, facial numbness and ptosis were caused by the tumor compressing the optic nerve, abducens nerve, trigeminal nerve and/or oculomotor nerve. Limb numbness was caused by the tumor compressing the medulla oblongata.

Preoperative imaging indicated that all lesions were located in the skull base. Details of tumor invasion sites are shown in Table 1 . The range of tumor invasion included the anterior skull base (2/17), sella (9/17), parasellar region (7/17), middle cranial fossa (4/17), medulla oblongata (1/17), and slope (3/17). The average tumor volume was 51.55 cm³ (range, 2.13–229.15 cm³). Cranial CT scans of the 13 patients showed clear boundaries between lesions and brain tissue, and calcification foci were observed in all lesions. We found patchy calcification in 10 patients, marginal calcification in 3, bone destruction in 2, and hypodensity shadow in 1. Additional CTA or MRA was performed in five patients to determine the relationships between tumors and surrounding blood vessels. On MRI, lesions were clearly demarcated from the surrounding brain tissue, and there were mixed signals with isometric or long T1 and long T2 signals, but the calcified part showed long T1 and short T2 signals. MRI enhancement showed uneven enhancement. According to preoperative-imaging findings, one case was diagnosed with craniopharyngioma, one with chordoma, one with schwannoma, two with cavernous hemangioma, three with pituitary tumor, four with meningioma, and five with chondroma ( Table 1 ).

We selected surgical approaches based on tumor size and location with minimal damage to brain tissue ( Table 2 ). Tumors were removed by EETA/EEETA in nine cases and craniotomy in eight. The surgical approaches for craniotomy mainly involved the posterior midline approach (1/17), orbitozygomatic approach (3/17), pterion approach (2/17), or subfrontal approach (2/17). Intraoperatively, tumors were found to be tightly adherent to the ICA (11/17) and to cranial nerves (CNs) II (6/17), III (6/17), V (4/17), and VI (4/17). Most tumors were located in the epidural space, but some had eroded into the subdural space. All lesions were closely associated with the skull and were milky white in color, and most clearly contained cartilage-like debris.

In this group, 10 cases were completely resected, 5 were subtotally resected, and 2 were partially resected ( Table 2 ). During assessment 1 week after surgery, facial and limb numbness on the right had disappeared in one case, facial numbness had improved in one case, but the patient had no improvement in visual acuity. Dizziness was relieved in one case, visual acuity had improved in three cases, headache was relieved in four cases, sexual function had no improvement in one case, and diplopia was unimproved in three cases.

Postoperative complications mainly included diplopia combined with right-limb hemiparesis (1/17), diplopia (2/17), eyelid ptosis (1/17), pulmonary infection (1/17), subcutaneous hydrops (1/17), cerebrospinal fluid leakage (CSFL; 1/17), and intracranial infection (ICI; 3/17; Tables 1 , 3 ). ICI was cured by continuous lumbar drainage and antibacterial treatment, and CSFL was cured by bedrest and continuous lumbar-cistern drainage. Subcutaneous fluid was cured by puncture and pressure bandaging. The case of pulmonary infection was cured by physical intervention and antibiotic treatment.

Pathological examination showed that the tumors consisted of differentiated cartilages, with or without varying amounts of ossification, calcification, fibrogenesis, or mucus formation. Pathology showed a trend of chondrosarcoma transformation in three cases. In order to distinguish chondroma from other tumors, immunohistochemistry (IHC) was performed in 10 cases and showed positive for S100 and Vimentin in 10 cases, Ki-67 ≤ 3% in 3 and Ki-67 >3% in 2, positive for SOX-9 in 1, and negative for epithelial-membrane antigen (EMA), cytokeratin (CK), CK18, CD34 and glial fibrillary acidic protein (GFAP) in 10 cases.

In this study, we evaluated clinical symptoms and imaging during outpatient visits or over the phone at 3 months, 6 months, and every year for 5 years after 6 months. Follow-up duration range was 3–60 months (average, 40.41 months). One patient received gamma knife treatment 3 months after surgery, while the remaining patients received no adjuvant therapy. Two patients died due to progression of tumor recurrence (including the patient who received postoperative radiotherapy). None of the remaining patients had tumor recurrence.

Seventy-three duplicate records were removed through database search, leaving a total of 1253 records available for title and abstract checking. Eighty-nine full-text articles were evaluated. Finally, 85 articles were included for analysis ( Figure 1 ).

All articles were predominantly case reports and included 164 cases. The proportion of ICDs was found to be approximately equal in men (n=79) and women (n=85), with a minimum age of 8 years and a maximum age of 73 years (mean 32.6 years). The clinical manifestations of ICDs mainly included headache (n=62), oculomotor paralysis (n=24), abducent paralysis (n=18), and epilepsy (n=18). ICDs primarily appeared on MRIs as irregular long T2 signal with punctate low T2 signal and a few of them containing long T1 and long T2 signals (7, 8). CT scans showed calcification in almost all tumors. ICDs were mainly distributed in skull base (n=87), convexity (n=30), intraparenchymal (n=25), falx cerebri (n=14), and others (n=8). The majority of cases were misdiagnosed as meningioma or low-grade glioma before surgery. All patients underwent surgical treatment, and almost all ICDs in convexity, intraparenchymal, and falx cerebri were gross-totally removed.

All chondromas in skull base were removed by craniotomy except for seven cases in which chondromas were removed by EETA, including GTR (n=18), STR (n=34), and PR (n=31). Extent of tumor resection was not recorded in four cases of chondromas in skull base, and one case died during hospitalization after craniotomy. Nine of all postoperative patients received radiotherapy.

On histopathology, ICDs usually presented as irregularly sized mature chondrocytes with calcification and fibrous components. Based on IHC, S-100 and vimentin were positive, and cytokeratin, epithelial membrane antigen, and Ki67 were negative.

A 50-year-old female presented with left-facial and right–lower-extremity numbness combined with choking and difficulty swallowing that had lasted 3 years. Physical examination showed hypoesthesia in the right-lower extremity; the rest of the neurological examination was negative. Head CT indicated a high-density shadow (about 2.5 cm × 1.7 cm) on the left side of the medulla oblongata, reaching up to the bottom of the fourth ventricle, protruding down into the foramina magnum, and displacing the medulla oblongata to the right due to compression. Cranial MRI scan showed isotonic T1 signal with mixed long T2 signal and speckled short T2 signal changes within the tumor ( Figure 2A ); the enhanced scan showed inhomogeneous enhancement ( Figure 2B ). Preoperative diagnosis was chordoma. A posterior midline approach was performed intraoperatively. The lesion was white, hard as bone and mostly located in the epidural space, with some erosions protruding into the dura and closely adhering to the medulla oblongata. We carefully separated the tumor along the medulla oblongata with a stripper, and total tumor resection was achieved under microscopic guidance. Postoperative MRI suggested there was no residual tumor in the surgical area ( Figure 2C ). The pathological diagnosis was chondroma ( Figure 2D ); ICH showed Vimentin and S-100 stain was positive, but Ki-67, CK, CK18, EMA, CD34 and GFAP stain was negative. After surgery, the numbness on the left side of the patient’s face and in her right-lower extremity disappeared, as did her coughing on swallowing. There was no tumor recurrence during 60 months’ follow-ups.

A 15-year-old female was admitted with diplopia that had lasted 6 months. She had a history of left–upper-extremity and right–lower-extremity fractures. On physical examination, the left eye was limited in abduction; the remaining neurological examination was negative. Cranial CT scan showed abnormal mixed density in the sellar and parasellar regions ( Figure 3A ). Cranial MRI scan suggested a long T1 signal and a long T2 signal in the parasellar region and the anterior and lateral pons, and multiple patches of short T1 and short T2 signal were found in the lesion ( Figures 3B, C ). After enhancement, the lesion showed heterogeneous enhancement ( Figure 3D ). The preoperative diagnosis was chondroma. Surgery was performed via the EEETA, creating the nasal septal mucoperiosteal flap first. Then, a grinding drill was used to remove the bone, including the anterior wall of sphenoid sinus, sellar floor, and part of slope and the anterior clinoid process, in order to expose the left cavernous sinus. Neuronavigation and Doppler helped clarify the tumor boundary, the location of the ICA and the extent of bone abrasion. We resected a portion of the tumor using a 0° scope and then probed toward the left petrous apex to resect the remaining tumor using a 30° scope. Neuronavigation aided in removing tumors adjacent to the anterior and lateral brainstem. Intraoperatively, the lesion was found to be grayish red in color, hard in texture, located in the epidural and partially closely adherent to the dura mater ( Figures 3E–H ). Postoperative MRI suggested GTR of the tumor ( Figures 3I–K ). On histopathological examination, the tumor was diagnosed as chondroma ( Figure 3L ); Vimentin ( Figure 3M ), Ki-67 (6%) ( Figure 3N ), SOX-9 ( Figure 3O ), and S-100 ( Figure 3P ) stained positive, but CK, CK18, EMA, CD34 and GFAP stained negative. The patient had no immediate improvement in diplopia or left-eye abduction restriction after surgery. The diplopia improved 3 months after surgery. No recurrence was observed during 60 months’ follow-ups.

A 24-year-old female had hypopsia in her left eye for 2 months. On physical examination, visual acuity of the left eye decreased; the remaining neurological examination was negative. Cranial MRI showed the tumor with irregular long T2 signal and dotted low T2 signal ( Figure 4A ), and the enhanced scan showed inhomogeneous enhancement ( Figure 4C ). The tumor communicated intracranially and extracranially. Magnetic resonance angiography (MRA) showed compression and displacement of the right ICA of the petrous and the cavernous segments ( Figure 4B ). Preoperative diagnosis was schwannoma. Surgery was performed by the EEETA, creating the septal mucoperiosteal flap first. After removing the bone of the anterior wall of sphenoid sinus and the posterior part of nasal septum, optic nerve, and ICA prominence, clival recess was exposed. Then, upon continuously grinding the petrous bone, the dura remained intact and then was incised by a hook knife. Hard white tumor tissue, partially protruding into the subdural space, was visible and was resected piecemeal ( Figures 4G–I ). Postoperatively, the patient had complications of limited abduction of the right eye and diplopia, and no improvement in left-eye visual acuity. Postoperative MRI suggested GTR of the tumor ( Figures 4D–F ). On histopathological examination, the tumor was diagnosed as chondroma ( Figure 4J ); Vimentin ( Figure 4K ), S-100 ( Figure 4L ) and Ki-67(2%) stained positive, but EMA, CK, CK18, CD34, SOX-9 and GFAP stained negative by ICH. Diplopia, right-eye abduction restriction and visual acuity improved 3 months after surgery. Six months after surgery, diplopia and right-eye abduction restriction had returned to normal. There was no recurrence during 24 months’ follow-ups.

This study was designed retrospectively and could not avoid the common biases that could have been avoided in controlled studies. In our study, the average time of follow-up was 40.41 months, which might require a longer duration to observe whether there was a recurrence of tumor after surgery. Additionally, due to the extremely low incidence of ICDs, further statistical evaluation was not done owing to a small number of cases.