Chordomas represent less than 1% of intracranial tumors and 3% of primary bone tumors, with an incidence of 0.08 per 100,000, peaking in the sixth decade, with a 2:1 male predominance.¹ Chordomas originate from embryological remnants of the notochord known as ecchordoses, found within the bony craniospinal axis. Clinically, chordomas concentrate around the clivus and craniocervical junction (32%) and the sacrococcygeal region (29%), with remainder distributed along the rest of the spinal axis (33%).² Although classified as a low-grade tumor, chordomas are associated with high morbidity and mortality and frequent local recurrence with invasion of surrounding tissues.

Clinical presentation is determined by the location of the lesion, direction of expansion, and rate of growth ( ▶ Fig. 1.1). Skull base chordomas ordinarily arise from the clivus in the midline and often present with symptoms related to expansion of the clivus, with stretching of the clival dura resulting in sixth nerve palsy with diplopia on lateral gaze. As the tumor extends to involve the sellar and suprasellar regions, symptoms of pituitary insufficiency and chiasmatic compression are common.³ Parasellar extension and invasion may produce features of cavernous sinus syndrome. Midclival lesions can present with signs and symptoms of pontine compression, and prominent extension into one side or the other can mimic cerebellopontine angle lesions. Lower clival tumors arising from the basion can compress the medulla and cause lower cranial nerve palsies and have been known to result in sudden death.⁴ Sacral chordomas present with local mass effect with or without symptoms of neural involvement. Low back pain is common, and radicular involvement leads to bladder and bowel dysfunction as well as pain and weakness in the lower extremities.

The primary management of chordomas is surgery, and the development of the surgical management for chordomas mirrors the history of neurosurgery over the past century. Insight into pathogenesis of the disease combined with the advent of modern imaging modalities and minimally invasive surgical techniques has improved outcomes, and there are emerging roles for adjuvant radiotherapy and stereotactic radiosurgery, as well as molecularly driven therapies.⁵

1.2 Origins of Chordomas

Virchow first characterized small, soft, jelly-like tissues arising from the synchondrosis spheno-occipitalis in the 1850s.⁶ He observed large, vacuolated, and plantlike cells, which he described as physaliphorous, derived from the Latin term for vacuoles or containing bubbles. Indeed, physaliphorous features on histology have become virtually pathognomonic for chordomas. In 1857, Luschka and others made similar observations and reported mucoid and gelatinous tumors arising from the clivus.⁷ These early studies suggested that these lesions were abnormal growths of cartilaginous origin, hence his term “ecchondrosis physaliphora spheno-occipitalis.” Hasse⁸ and Zenker⁹ confirmed these findings and referred to the growths as “gelatinous tumors of the clivus Blumenbachii.” The following year, based on anatomical and embryological observations, Müller postulated that these growths were associated with the notochord and that they should be termed “ecchordosis physaliphora” (originating from the Chorda dorsalis); he fundamentally rejected the notion that that these lesions were of cartilaginous origin.¹⁰

In 1864, Klebs gave the first description of pontine compression from a skull base chordoma.¹¹ It was not until 1894, however, that Ribbert first proposed the term “chordoma” to describe experimental lesions in rabbits that occurred following the puncture of ligaments surrounding intervertebral discs and resembled ecchordoses. The relationship between these small and asymptomatic cellular nodules, presumably composed of nucleus pulposus, and the large invasive chordomas was not yet clearly understood. Based on autopsy studies in humans, small ecchordoses along the dorsal clivus are present in 2% of the population, and benign notochordal tumors of the clivus are recognized to be present in 11.5% of autopsies.¹² It remained unknown whether these small benign ecchordoses progressed to classic chordomas, and if so, the mechanism of transformation ( ▶ Fig. 1.2).

Ribbert investigated anterior intervertebral ligament puncture in rabbits and ultimately produced an animal model of chordoma formation, allowing him to classify them as developmental tumors. In 1904, he published his opinion on the origin of chordomas, and in his laboratory, Fischer and Steiner confirmed Müller’s theory by inducing malignant chordoma in rabbits 3 years later.¹³ Subsequently, the association with the notochord would be made. In 1909, Linck established criteria for the histopathologic diagnosis of chordoma and described mucus formation, presence of physaliphorous cells, lobular arrangement, nuclear vacuolation, and resemblance to notochordal tissue as characteristic findings.¹⁴ That same year, Harvey Cushing, then at the Johns Hopkins Hospital, carried out the first successful resection of a skull base chordoma, which he reported in 1912.¹⁵

Surgeons and scientists continued to study chordomas, and their efforts eventually translated to patient care. In 1952, Congdon redemonstrated Ribbert’s rabbit chordoma model.¹⁶ Zulch, working in Germany in 1956, was able to distinguish two separate types of chordoma: slowly growing benign lesions and rapidly growing malignant lesions. Astutely, Zulch also proposed that chordomas arose where notochord tissue had no cartilaginous envelope.¹⁷ The “chondroid” subtype of chordoma was described by Heffelfinger and colleagues in 1973, after investigating the pathology of 155 cases of chordoma treated at the Mayo Clinic from 1910 to 1973.¹⁸

1.3 Pathogenesis and Pathology

Histologically, chordomas are divided into three subtypes: conventional, chondroid, and dedifferentiated. Conventional chordomas are composed of nests and cords of epithelial-appearing cells and range from dense eosinophilic cells in a mucinous matrix to cells with large cytoplasmic vacuoles—the classic physaliphorous appearance. Chondroid chordomas exhibit cellular differentiation into cartilaginous or bony tissue and are associated with a better prognosis.¹⁸ Infrequently, chordomas may display atypia and increased proliferative index and are classified as dedifferentiated, although distant metastasis is unusual.

Chordomas and chondrosarcomas are often considered together because of their similar clinical presentation, sites of occurrence, radiographic appearance, and surgical management.¹⁹ The two, however, are distinct pathologic entities, with different origins and pathogenesis. In contrast to the notochordal origin of chordomas, chondrosarcomas are cartilaginous tumors of mesodermal origin. Histologically, because of their overlapping features, chordomas may be difficult to differentiate from chondrosarcomas. Chondrosarcomas carry a more favorable prognosis. Chordomas are distinguished from chondrosarcomas by their expression of epithelial antigens, such as epithelial membrane antigen (EMA) and cytokeratin.²⁰ In contrast, chondrosarcomas do not express epithelial antigens. The question has also been raised as to whether chondroid chordomas are in fact a subtype of chondrosarcoma and hence account for the more favorable outcomes observed. Using epithelial markers, chondroid chordomas have been shown to be true chordomas and are not related to chondrosarcomas.

Molecular evidence further supports the notochordal origin of chordomas. Brachyury is a transcription factor encoded by the T gene, and it is an important regulator of notochord development. It is exclusively expressed in chordomas and absent in chondrosarcomas and other connective tissue tumors²¹ and thus provides a diagnostic marker. Genetic screening in rare chordoma families suggests that brachyury expression may contribute to the pathogenesis of chordomas. Using DNA array comparative genomic hybridization, which detects chromosomal aberrations in the form of copy number imbalances, a duplication of the 6q27 T gene–containing region was identified in four chordoma families.²² T gene duplication in familial chordomas and its amplification in sporadic cases certainly suggest a strong association.²³ It remains to be determined whether brachyury has a causal role in chordoma formation, as it has now been linked to a variety of neoplastic processes.

1.4 Surgical Management: A Historical Perspective

In 1909, Harvey Cushing was the first surgeon to successfully remove a cranial chordoma. Published in The Pituitary Body and Its Disorders in 1912, Cushing described his surgical approach to a clival chordoma. In case XVII, a 35-year-old man presented with headache, progressive visual loss, left oculomotor palsy, and hypopituitarism ( ▶ Fig. 1.3). His radiograph showed “practically complete absorption of all sellar landmarks.” Believing it to be a lesion affecting primarily the sellar region, Cushing performed a transnasal, transsphenoidal procedure and decompressed the sellar base using von Eiselsberg’s method. He noted the “obliteration of the sphenoid cells” and “complete pressure atrophy of the sellar base.” “Some large fragments of a friable growth” were removed, and postoperatively, the patient reported improvement in severity of headache and visual fields. Histologically, “the tumor consisted of embryonic cartilage containing a few bone cells with myxomatous and connective tissues.” William Welch, the preeminent pathologist, suggested that the lesion was a “mixed tumor from a congenital anlage,” and the tumor was erroneously labeled a teratoma in Cushing’s monograph. A subsequent review of the histology by Bailey and Bagdasar suggested that the tumor contained typical features of a chordoma.²⁴ The patient was readmitted 6 months postoperatively with worsening of symptoms and died following a repeat operation. At autopsy, a 5-cm round lesion was noted to displace the hypophysis and infundibulum anteriorly, whereas the midbrain and pons were displaced posteriorly ( ▶ Fig. 1.3). Although the relation of the tumor to the clivus was not directly discussed, the anatomical location of the lesion and the histologic features were consistent with a diagnosis of clival chordoma. Notably, Cushing lamented: “from a surgical standpoint, complete extirpation of such a growth is unthinkable. Palliative measures alone must be resorted to.” He understood that although complete excision could not be achieved, repeat surgical debulking of the tumor could provide local disease control as a means of ameliorating symptoms and improving the patient’s quality of life.⁵

Chordoma as a distinct clinical entity was formally recognized in the medical literature in the 1920s.²⁵ Stewart reported an early series of spheno-occipital chordoma at the University of Leeds.²⁶ The patients had tumors that localized to the extremities of the primitive notochord in the spheno-occipital (clival) and sacrococcygeal regions. Histologically, the tumor specimens contained the uniform mucinous vacuolated appearance of Virchow’s physaliphorous cells. Even at this early stage, clival chordomas were recognized to present with symptoms of pituitary dysfunction, visual loss, and pressure effects on the cranial nerves and the brainstem ( ▶ Fig. 1.4).

1.4.1 Early Surgical Experience

Surgeons since Cushing’s time have echoed his view that extirpation of cranial chordomas poses formidable challenges, given the complex anatomy of the clival region and surrounding vital structures. With continued progress in surgical technique, however, a more radical approach to the skull base would eventually become possible.²⁷ In 1952, Dahlin and MacCarty at the Mayo Clinic reported a large series of 59 cases of chordomas.²⁸ Significantly, half of these cases involved the sacrum, and the high incidence of recurrence following partial resection prompted the development of radical surgical techniques. To this end, an en bloc removal of the coccyx and most of the sacrum, together with tumor, was performed, and the principle of en bloc resection for the management of chordomas was initiated.²⁹ Also in 1952, Poppen and King described different surgical approaches, including subfrontal, frontal–temporal, and suboccipital approaches, for the resection of clival chordomas. Their outcomes were poor, as five of seven patients died in the immediate postoperative period.³⁰

Despite major progress in microneurosurgery in the 1960s, a hundred years after the first description of chordomas, surgical treatment still produced disappointing results. By the time the tumor becomes clinically apparent, complete tumor removal can be hazardous, although debulking surgery by suction of the soft mucinous tumor contents may alleviate symptoms for varying lengths of time.³¹ In fact, some large tumors that presented with involvement of the sphenoid and maxillary sinuses, nasal cavity, nasopharynx, and surrounding structures were diagnosed and managed by otorhinolaryngologists. Initial neurosurgical experience was based on clinical localization and imaging modalities, such as plain-film X-rays and ventriculograms.³²

In a review of 129 cases from the literature in 1975, Yasargil reported disappointing outcomes for the majority of patients.³³ Surgical mortality and morbidity was high, and 38 (27%) patients died within 3 months of surgery, including 24 (17.5%) perioperatively.³³ One-year survival was 40%, and only 5 (3.5%) patients survived 10 years.³³ Given the overall poor survival, patients may have had chondrosarcomas rather than chordomas.

1.5 Improving Outcomes: Modern Imaging, Microsurgery, and the Role of Endoscopy

The 1980s ushered in a new era in neuroimaging, with the advent of the magnetic resonance imaging (MRI), and this in turn helped to reinvigorate the notion of gross total resection (GTR) of chordomas.³⁴ In the early 1990s, Yasargil analyzed 25 consecutive chordoma cases from the preceding 18 years, and reported 10.5% reduction in surgical mortality at 3 months.³⁵ One-year survival doubled to 84%, with 47% surviving for 4 years or more, and 21% of patients remained alive after 10 years.³⁵ The Mayo Clinic reported similarly improved survival rates for 51 patients treated from 1960 to 1984, with 5- and 10-year survival rates of 51% and 35%, respectively.³⁶ The improvement in overall survival is also documented by analysis of the National Cancer Institute Surveillance, Epidemiology, and End Results (SEER) database ( ▶ Table 1.1).³⁷ A more recent study suggests that overall survival in cranial chordomas has continued to increase, with one center reporting 93% 5-year survival for patients treated between 2000 and 2011.³⁸ Several factors may account for such improvements in survival, and will be discussed individually.

1.5.1 Microscope-Assisted Surgery

Since the 1970s, widespread adoption of computed tomography (CT) and MRI for preoperative planning and the advent of microscope-assisted transcranial techniques with en bloc or gross total resection (GTR) where possible have significantly improved surgical outcome and increased recurrence-free survival. The principle of en bloc resection has been successfully applied to chordomas of the skull base ( ▶ Fig. 1.5), spine, and sacrum with good results. In 52 patients with sacral chordomas followed for a mean of 7.8 years at the Mayo Clinic, GTR with wide margins conferred significant benefits in terms of recurrence-free survival and overall survival, compared with those who underwent partial resection.^39,​40

Related posts:

Surgical Decision Making in Chordomas Molecular Imaging of Chordomas Radiation Treatment for Chordomas Craniovertebral Reconstruction after Chordoma Resection The Extended Petrosal Middle Fossa Approach Chordomas: A Personal Perspective

Stay updated, free articles. Join our Telegram channel

Join