Lysis of type-I collagen by squamous carcinomas of the head and neck has been studied in freshly excised tissues, xenografts and established cell lines. Investigations with 35 freshly excised tumours showed only low levels of active and total collagenase in both carcinomas and controls. A difference became apparent when the tissues were set up in explant organ culture where a significant (p less than 0.05) increase in total collagenase was found in 13/19 tumours compared with paired control tissues over a 4-week culture period. Two xenografts showed little capacity to lyse collagen in vitro and there was only limited evidence of an increase in total collagenase after explantation and growth in organ culture. Twenty tumour cell lines showed low levels of active collagenase. Total collagenase levels were significantly increased (p less than 0.05) in 4 of the cell lines derived from cancers of the tongue; this activity was sustained in subsequent passages. Six control fibroblastoid cell lines also showed low levels of active collagenase. Levels of total collagenase were consistently high, but this activity was transient and declined in subsequent passages. Co-cultivation experiments with II tumour-cell lines and 5 fibroblastoid cell lines showed some enhanced, synergistic destruction of collagen. Parallel experiments with supernatant media from the carcinoma and fibroblastoid lines showed no enhancement, indicating that intact carcinoma cells and fibroblastoid cells are required for synergistic collagenolysis to take place.

The accuracy of standard clinical and radiologic methods in detecting invasion of the laryngeal framework or the presence of a transglottic tumor was examined by comparing clinical and pathologic findings in a series of 50 randomly selected laryngectomies for squamous carcinoma. Forty-two of the patients had received radiation therapy as primary treatment. The presence of pain referred to the ear immediately before laryngectomy indicated invasion of the laryngeal framework or spread into the extralaryngeal soft tissues in 11 of 12 patients. Transglottic tumors were correctly identified in 10 of 14 patients from laryngeal tomograms, but seven additional tumors were incorrectly designated as transglottic. Both transglottic tumors and infraglottic extension were seriously underdiagnosed by direct laryngoscopy. Framework invasion was seen to be a feature of transglottic tumors 2 cm or more in diameter. The morphology of this process is described. Limited follow-up has already shown a pattern of early recurrence, frequently as cervical node metastases, that is significantly more common in patients with framework invasion.

Fragments of VX2 squamous carcinoma were transplanted into the femoral triangles of rabbits and the growth patterns of tumour invasion were compared in intact femoral arteries, ligated femoral arteries and in femoral arteriovenous (AV) anastomoses. Intact arteries were resistant to tumour; ligated arteries were invaded by tumour and, in most instances, destroyed; AV anastomoses were also invaded but some mural structures remained intact. The relative resistance of systemic arteries to neoplastic invasion appears to be due to a combination of the normal structure of the arterial wall and the normal dynamics of the arterial circulation. Infiltration is facilitated if arterial perfusion falls and/or the normal structure of the arterial wall is modified.