BACKGROUND An accurate pretherapeutic staging of laryngeal carcinoma is required for most treatment planning as well as for evaluation and comparison of the results of different treatment modalities. Neoplastic invasion of the laryngeal cartilage may have important therapeutic implications. To our knowledge, no data are available comparing the impact of endoscopic examination, computed tomography (CT), and magnetic resonance (MR) imaging on pretherapeutic staging accuracy. The purpose of our study was to determine which imaging should be used as an adjunct to other clinical examinations in the pretherapeutic staging of laryngeal carcinoma. METHODS In this study, 40 consecutive patients with neoplasms of the larynx, who were treated surgically, were included in a prospective pretherapeutic staging protocol that included indirect laryngoscopy, direct microlaryngoscopy, contrast-enhanced CT, and gadolinium-diethylenetriamine pentaacetic acid-enhanced MR imaging at 1.5 Tesla. The surgical specimens were cut in whole-organ slices parallel to the plane of the axial CT and MR images. The histologic findings were compared with the clinical findings including the CT and MR images. The impact of each diagnostic method on pretherapeutic staging was analyzed. RESULTS Clinical/endoscopic evaluation failed to correctly stage 17 tumors due to invasion of the paraglottic space (1 tumor), preepiglottic space (2 tumors), and extralaryngeal soft tissues (14 tumors), resulting in a pretherapeutic staging accuracy of 57.5%. Neoplastic invasion of cartilage was present in 28 patients and absent in 12 patients. Although MR imaging was more sensitive in detecting neoplastic invasion of cartilage than CT (94% vs. 67%; P = 0.001), MR imaging was less specific than CT (74% vs. 87%; P = 0.007). There was no difference between the overall accuracy of CT and MR imaging in detecting neoplastic invasion of cartilage (80% vs. 82%). The accuracy of combined clinical/endoscopic examination and CT staging was 80% and the accuracy of combined clinical/endoscopic examination and MR imaging staging was 87.5%; the difference was not statistically significant. CONCLUSIONS Clinical/endoscopic examination alone failed to identify tumor invasion of the laryngeal cartilages and of the extralaryngeal soft tissues, resulting in a low staging accuracy (57.5%). Many pT4 (according to the International Union against Cancer TNM Staging System) tumors were clinically unrecognized. The combination of clinical/endoscopic evaluation and an additional radiologic examination, either CT or MR imaging, resulted in significantly improved staging accuracy (80% vs. 87.5%). MR imaging is significantly more sensitive but less specific than CT in detecting neoplastic cartilage invasion. Therefore, MR imaging tends to overestimate neoplastic cartilage invasion and may result in overtreatment, whereas CT tends to underestimate neoplastic cartilage invasion and may lead to inadequate therapy.