Introduction: Internal thoracic vessels have been gaining increasing acceptance for primary and delayed reconstruction in mastectomy patients. Use of the internal thoracic vessels as recipients in breast reconstruction was first reported in the early 1980s.They were initially used at the level of the fifth costal cartilage, resulting in problematic discrepancies in caliber. Their use avoids opening the axilla and may lessen the risk of lymphedema and nerve and vessel injury. These vessels have the further advantage of consistant location, are of large caliber and permit easy medialization of the TRAM flap when dissected with an adequate pedicle. Disadvantages of the traditional harvesting technique include the necessity of resecting costal cartilage with the resultant chest wall defect, limited pedicle length and postoperative chest wall discomfort. We report a new operative technique and our preliminary experience in a group of breast cancer patients who underwent TRAM and Superior Gluteal flap breast reconstruction using internal thoracic vessel pedicles that were endoscopically harvested with robotic assistance. The purpose of the study was to gain insight into how endoscopic recipient vessel harvesting affects surgical complications and outcomes. Methods: The prospective TRAM flap database at Cleveland Clinic Florida was searched for breast cancer patients who underwent flap reconstruction between March, 2001 and November, 2003 after having undergone endoscopic internal thoracic vessel harvesting with robotic assistance. Patients were placed in the right lateral decubitus position and with ipsilateral lung collapse and the internal thoracic pedicle harvested endoscopically from the 1st to 6th intercostal space using a harmonic scalpel. In all cases a voice-activated robotic manipulator was used to guide a 5mm 30 degree endoscope. Two of the three 5mm ports were placed through the mastectomy incision. An intercostal incision was created internally and the pedicle externalized under endoscopic visualization after distal vessel control. Demographic data, patient factors known to affect TRAM flap outcome, and radiotherapy data were collected from a detailed chart review. The patients' records were also reviewed for information on all recipient vessels dissected. Outcomes including total flap loss, partial flap loss, vascular complications, fat necrosis, and arm lymphedema were also determined. Results: 22 patients underwent tissue flap breast reconstruction during the study period. Average patient age was 53.7 years. 11 (50%) of patients had preoperative radiation treatment, of these 3 (14%) had radiation skin damage. 13 patients had previous axillary dissection. Endoscopic dissection times progressively decreased with experience from a maximum of 180 minutes to 45 minutes. Pedicle length averaged 6.9cm. Recipient artery size averaged 2.52mm, recipient vein size averaged 2.3mm. Over 80 percent of the time there were two accompanying veins, with the venous junction most commonly at the level of the third or fourth intercostal space. There was excellent arterial flow in all harvested pedicles. There were two flap failures. One failure was due to vein injury during harvesting. Two patients had transient brachial plexus injuries, and one patient developed a postoperative pneumothorax.

Conclusion: Endoscopic internal thoracic vessel harvesting was successful in 21/22 (96%) patients. Advantages of this technique include limitation of the chest wall defect, limited postoperative chest discomfort, easy recipient vessel preparation and maximal “freedom of flap movement”. Drawbacks to the technique include the risk of brachial plexus injury, the learning curve for the endoscopic harvesting technique, the requirement for additional instrumentation and a chest tube postoperatively.