Sternal cleft is a chest wall malformation that can expose mediastinal viscera and vessels to injuries. It can be classified into two forms, complete and partial. Its etiology and incidence are unknown and it is often associated with other defects. The aim of this article is to review the literature and report our experience with this rare anomaly, focusing on clinical presentation and management. We reviewed the English written literature about sternal cleft and collected the clinical data of all the published series. We present seven new cases that we have observed and treated since 1999. Literature reports 51 series including 86 patients, more frequently female (62%) and affected with partial superior form (67%). Sternal cleft is often asymptomatic (74%) and associated with other defects (72%). Surgical treatments include primary closure (73%), bone graft interposition (10%), prosthetic closure (7%), and muscle flap interposition (3%). In our series, primary closure was possible in four cases, while in three cases we placed a prosthesis. Five patients had associated defects and two were affected with PHACES (posterior fossa abnormalities, hemangiomas, arterial lesions, cardiac abnormalities/aortic coarctation, abnormalities of the eye, and sternum defects) syndrome. We report for the first time the association of sternal cleft with connectival nevi in three of our patients. At follow-up, we observed no major complication or recurrences. Although primary closure is the preferred option and should be performed in the neonatal period, the use of prostheses warrants good results as well. Prior to treatment, associated defects and syndromes should be excluded.

Purpose: The International Neuroblastoma Response Criteria (INRC) require serial measurements of primary tumors in three dimensions, whereas the Response Evaluation Criteria in Solid Tumors (RECIST) require measurement in one dimension. This study was conducted to identify the preferred method of primary tumor response assessment for use in revised INRC. Patients and Methods: Patients younger than 20 years with high-risk neuroblastoma were eligible if they were diagnosed between 2000 and 2012 and if three primary tumor measurements (antero-posterior, width, cranio-caudal) were recorded at least twice before resection. Responses were defined as ≥ 30% reduction in longest dimension as per RECIST, ≥ 50% reduction in volume as per INRC, or ≥ 65% reduction in volume. Results: Three-year event-free survival for all patients (N = 229) was 44% and overall survival was 58%. The sensitivity of both volume response measures (ability to detect responses in patients who survived) exceeded the sensitivity of the single dimension measure, but the specificity of all response measures (ability to identify lack of response in patients who later died) was low. In multivariable analyses, none of the response measures studied was predictive of outcome, and none was predictive of the extent of resection. Conclusion: None of the methods of primary tumor response assessment was predictive of outcome. Measurement of three dimensions followed by calculation of resultant volume is more complex than measurement of a single dimension. Primary tumor response in children with high-risk neuroblastoma should therefore be evaluated in accordance with RECIST criteria, using the single longest dimension.