Search

Introduction: A prospective comparative study with pathology was performed at the National Cancer Institute, Milan, to assess the clinical value of Computed Tomography (CT) and endoscopic ultrasound (EUS) for nodal staging in lung cancer., Material and Methods: In three years, 71 patients with histological diagnosis of non-small-cell lung cancer were operated on. They underwent CT and EUS examinations to identify mediastinal lymphadenopathies after major nodal involvement had been excluded by chest X-ray. Diagnostic staging was completed in two weeks prior to treatment. Patients received complete tumor removal and radical lymphadenectomy (55 patients), invasive staging with node resection and sampling (11 patients), or mediastinoscopy (5 patients). Blinded interpretation of CT alone, EUS alone, and CT and EUS together were performed, with systematic correlation of imaging findings and pathological results., Results: The frequency of mediastinal involvement was 42.2%. A total of 329 nodal stations were dissected or sampled and 755 lymph nodes were examined at histology. On a per-station basis, CT had greater sensitivity (74%) than EUS (56%), but EUS was more specific (83.4% vs 92.7%). The accuracy rates of the two techniques were similar (CT 81%, EUS 83%). A site by site analysis showed highest sensitivity (100%) in the lower right paratracheal nodes for CT, and in the superior left paratracheal and subcarinal nodes for EUS. When the EUS and CT images were studied together by specialists on a per-station basis, sensitivity, specificity, and accuracy increased to 85%., Conclusions: Endoscopic ultrasound should be part of the routine preoperative diagnostic approach to non-small-cell lung cancer, because of its high specificity. Results can be improved when EUS and CT are combined, which suggests that these imaging modalities should be used together in selected patients for the noninvasive staging of non-small-cell lung cancer to identify local lymphatic spread.

The records of 3,795 cases of malignant melanoma treated at the INT (Milan) from 1975 to 1992 were reviewed. Histologic confirmation was obtained in all cases. Thirty-one patients (0.82%) with solitary or multiple skeletal metastases were identified. The review of conventional films, tomograms, CT, MR and bone scintigraphy images enabled us to detect 120 single bone lesions. The X-ray features were divided into two groups according to typical and atypical skeletal lesions. Typical bone metastases are osteolytic (87.5%), with medullary origin (91.6%), and they cannot be distinguished from other osteolytic metastases on the basis of imaging criteria alone. Lesion growth causes cortical erosion and destruction (46.6%), pathologic fractures (22.5%) and soft tissue involvement (12.5%). Lytic areas usually have ill-defined margins. Clear-cut outline is an uncommon finding. Atypical skeletal metastases exhibit a mixed osteolytic-osteoblastic pattern (10%), which is hardly ever completely osteoblastic (2.5%). Other unusual metastatic patterns include intense trabecular rarefaction with no detectable single lesion (3.3%), the presence of a well-defined sclerotic rim and periosteal reaction (12.5%). Atypical growth may cause extensive cortical destruction and periosteal production resembling osteogenic osteosarcoma. The various imaging methods show that conventional radiology has relatively poor sensitivity because of anatomical reasons, while MRI is the most sensitive method to detect skeletal localizations. Treatment changes the radiologic patterns of the lesions: recalcification, sclerotic rim, periosteal reaction are common response patterns. Finally, in spite of the above limitations, conventional radiology remains the method of choice to assess lesion evolution during the follow-up.