I for a primary sarcoma. Because of dangers associated

Sternal Resection and Reconstruction Kamal A. Mansour, MD, Timothy M. Anderson, MD, and T. Roderick Hester, MD Division of Cardiothoracic Surzerv and Division of Plastic and Reconstructive Surgery, Joseph B. Whitehead Department of Surgery, Emory University Sch'od of Medicine, Atlanta, Georgia Twenty-one patients underwent sternal resection and reconstruction. Surgical indications included sternal infection in 9 patients, recurrent breast in 6, metastatic carcinoma from an unknown primary in 2, pectus excavatum in 2, and osteogenic sarcoma and eosinophilic granuloma in 1 each. Management included partial sternectomy in 10 patients (group 1) and complete sternectomy in 11 (group 2). Chest wall reconstruction was by various flaps and mesh repairs. Blood transfusions averaged 2 units in group 1 versus 5.5 units in group 2 (p = 0.02). Average number of days until extubation was 2.6 in group 1 versus 7.3 in group 2 (p = 0.04). Average number of intensive care unit days was 4.4 for group 1 versus 9.4 for group 2 (p = 0.03). The number of days until discharge was 14 days for group 1 versus 20 days for group 2. Complications occurred in 40% of group 1 and 82% of group 2 patients. Overall mortality was 9.5%. Sternal resection and reconstruction, particularly complete sternal resections, are a major undertaking with substantial morbidity. Using a multidisciplinary approach (cardiothoracic, plastic and reconstructive, critical care medicine, and infectious disease) and aggressive pulmonary support, acceptable cosmetic and functional results are possible. (Ann Thorac Surg 1993;55:838-43) n 1878, Holden [l] first described partial sternectomy I for a primary sarcoma. Because of dangers associated with entering the pleural spaces and inability to replace the entire sternum or stabilize the anterior chest, subsequent reports kept all or part of the sternum intact to lend stability to the chest wall [2, 31. The advent of antibiotics, positive-pressure ventilation, thoracic suction drainage, reliable synthetic substances for chest wall protection, and musculocutaneous flaps [4] all paved the way for the first cases of total sternal resections in 1959 by Brodin and Linden [5], followed in 1960 by Baue [3]. Today's techniques of modem reconstruction have further expanded indications for chest wall resection [6]. This retrospective study reviews our experience with both partial and complete sternal resections. Material and Methods We reviewed the records of 21 patients who underwent sternal resection and reconstruction from 1984 to 1991 at Emory University Hospital. Follow-up was by telephone conversation with the patient, the family, or the primary physician. Statistical analysis was by independent group t test. There were 11 female and 10 male patients, with an average age of 56 years (range, 6 to 76 years). Symptoms were present in 18 of the patients. Wound drainage was the most common complaint expressed by 10 patients, followed by pain in 9, dyspnea in 3, fever in 2, and recurrent infection in 1. Three patients were asymptomatic. Presented at the Thirty-ninth Annual Meeting of the Southern Thoracic Surgical Association, Wesley Chapel, FL, Nov 5-7, 1992. Address reprint requests to Dr Mansour, Cardiothoradc Surgery, The Emory Clinic, 1365 Clifton Rd, NE, Atlanta, GA 30322. The most common physical finding was a draining tract in 10 patients, followed by a mass in 9, tenderness in 7, ulceration in 4, erythema in 2, and pectus excavatum in 2. One patient had a normal physical examination. Results Diagnostic Evaluation Laboratory data before operation were evaluated including hematocrit and levels of alkaline phosphatase, total and direct bilirubin, serum glutamic oxaloacetic transaminase, and albumin. Alkaline phosphatase level was elevated secondary to metastatic disease from breast carcinoma in 2 patients, and from an unknown primary in a third. However, for the most part laboratory results were not helpful. The most commonly employed diagnostic test was the preoperative chest roentgenogram performed in all 21 patients. Either a mass or density was found in 5 patients on chest roentgenogram. Rib films revealed evidence of infection in 1 patient and demineralization in another. Computed tomographic scan (CT) of the chest was the most useful test, revealing a mass or density in 7 patients and bony destruction in 5 patients (some patients had both). Computed tomography delineated the degree of pectus excavatum in 1 patient and was unremarkable in only 1 of 11 patients. Bone or gallium scan showed increased sternal uptake in 4 of 5 patients. For patients with suspected tumor, fine-needle aspiration was diagnostic for malignancy in 4 of 5 patients. Indications The most common indication for operation was sternal infection, present in 9 patients, followed by recurrent breast in 6. There were 2 patients with metastatic carcinoma from an unknown primary tumor, 2 with 0 1993 by The Society of Thoracic Surgeons

Ann Thorac Surg 19!33;55:838-43 MANSOUR ET AL 839 Table I. Indications and Reconstruction for Patients Undergoing Partial Sternal Resection (group 1) Patient No. Indication Reconstruction 1 Infection Rectus abdominis 2 Infection flaps, VM Rectus abdominins 3 Infection flap Rectus abdominis 4 Infection flap Omentum, STSG, 5 Infection VM Primary dosure 6 Osteogenic sarcoma Pectoralis major 7 Adenosquamous, Unkl flaps, PM Pectoralis major 8 Adenocarcinoma, Unkl flaps, STSG Primary closure, PM 9 Recurrent breast Rectus abdominis 10 Eosinophilic granuloma flap, PM Pectoralis major flap, PM PM = Prolene mesh; STSG = split-thickness skin graft; Unkl = unknown primary tumor; VM = Vicryl mesh. pectus excavatum, and 1 each with osteogenic sarcoma and eosinophilic granuloma (Tables 1, 2). Of note, 5 of 6 patients with recurrent breast also had previously received irradiation of the chest wall. One patient had both osteomyelitis and radionecrosis of the sternum secondary to an infected upper mediastinal dissection wound for squamous cell carcinoma (SCCA) of the head and neck in a previously irradiated field. The other patient with recurrent SCCA of the larynx underwent previous irradiation, upper mediastinal exploration, and tracheos- Table 2. Indications and Reconstruction for Patients Undergoing Complete Sternal Resection (group 2) Patient No. Indication Reconstruction 1 Recurrent breast Rectus abdominis flaps, PM 2 Recurrent breast Latissimum dorsi flaps, PM, STSG 3 Recurrent breast Fasciocutaneous flap, PM 4 Recurrent breast Fasciocutaneous flap, STSG, PM 5 Recurrent breast Omentum, PM, STSG 6 Infection Pectoralis major flaps, STSG 7 Infection Pectoralis, rectus flap, PM, STSG 8 Infection Rectus muscle flaps 9 Infection Omentum, primary closure 10 Pectus excavatum Pectoralis major flap, PM 11 Pectus excavatum Pectoralis, rectus, omentum, PM I'M = Prolene mesh; STSG = split-thickness skin graft. tomy before osteomyelitis of the upper sternum developed. The surgically created soft tissue defects ranged in size from 45 to 464 cm2, averaging 133.2 cm2 in group 1 and 241.3 cm2 in group 2 (p = 0.02). Surgical Resection Ten patients in group 1 received a partial sternectomy (Fig 1) to include portions of adjacent ribs ranging in number from 2 to 7 unilaterally or bilaterally. Of these, 1 patient also had an en bloc pericardiectomy and lung resection, and another required an aortic homograft for an infected aortic pledget. Eleven patients in group 2 underwent complete sternal resection, more than half of which were for local recurrence. Of these patients, 4 also required pericardiectomy, 2 needed lung resections, and 1 each needed tracheal resection or breast or thymus excision. The rationale of our approach for total sternectomy in patients with sternal infection after coronary bypass operation is to tackle the sternum from the periphery, leaving the critical point of sternal attachment to the heart and great vessels to the last. Cardiopulmonary bypass standby is made available, and the groin is prepared in the field. Through a midline incision, the skin, subcutaneous tissues, and pectoralis major muscle are dissected laterally on both sides. The right third costal cartilage is excised about 3 to 4 cm lateral to the sternum to avoid injuring the internal mammary artery. Once the pleural space is entered, sections of the second and first costal cartilages and the head of the clavicle are divided. Then the costal cartilages are divided distally down to the xiphoid process. The dissection is then carried across to the left side and all the costal cartilages and the head of the clavicle are divided in a similar fashion. Thus, the whole sternum and divided costal cartilages are freed Group1 1 Group 1 1 Group 1 1 Fig 1. Thoracic cage diagrams demonstrating complete sternal resection in 11 patients (group 2; upper left) and assorted partid sternal resection in 10 patients (group 1).

840 MANSOURETAL Ann Thorac Surg 1993;55838-43 from the chest wall. We then start from below and dissect the sternum off the underlying heart using the electrocautery. In patients with malignant disease, the line of resection is at least 2.5 cm beyond the margin of the tumor or the irradiated skin. Reconstruction Prosthetic materials were used in two-thirds of the patients (n = 14). Prolene mesh (Ethicon, Inc, Somerville, NJ) in 11 patients and Vicryl mesh (Ethicon, Inc) in 2 was doubled over and sutured to the adjacent ribs and fascia to cover the immediate surface of the skeletal defect. Methylmethacrylate was sandwiched between two layers of Prolene in 1 patient in an attempt to establish normal chest wall contour. Seven patients did not require prosthetic material. Of these, 6 patients had osteomyelitis, including 1 with concomitant radionecrosis. The seventh patient had adenosquamous carcinoma from an unknown primary tumor. Musculocutaneous flaps plus mesh were most commonly employed for closure of the defect (n = 9). In addition, musculocutaneous flaps with or without splitthickness skin graft were all that was necessary in 5 patients. Pedicled omentum was used in 3 instances. Transfusion Requirements Six of 10 patients in group 1 required blood transfusions postoperatively, compared with almost all patients (10111) in the complete resection group. One patient undergoing concomitant open heart operation was not included in these data. Of patients transfused, the average number of units given in group 1 was 2 f 2 units per hospital stay, compared with 5.5 & 3.6 units in group 2 patients (p = 0.02). Extubation, Number of Intensive Care Unit Days, and Discharge Interval Patients in group 1 had an average of only 2.6 -+ 2.9 days until extubation, and 3 were extubated in the recovery room. In group 2, the average number of days until extubation was 7.3? 5.8 (p = OM), and no one was extubated in the recovery room. The average stay in the intensive care unit was 4.4 * 2.9 days for group 1 compared with more than twice as long for group 2, averaging 9.4 k 5.9 days (p = 0.03). The number of postoperative days until discharge from the hospital was 14 & 7.2 days for group 1 compared with 20 f 6.9 days for group 2. Complications In group 1,3 patients had an early complication. One had wound cellulitis that resolved with antibiotics. Another had wound dehiscence, which healed by secondary intention in 6 weeks. The third had skin puckering relieved by releasing a Prolene stitch. The one late complication in group 1 was prolonged wound care, taking 3 months to heal. In group 2, 8 patients had one or more early complications. Pneumonia was most common, occurring in 4. patients, followed by wound dehiscence in 3. There was one bronchopleural cutaneous air leak that resolved spontaneously about 10 days postoperatively. Flap necrosis occurred in 1 patient, and wound revision was required in 1. Our one early death was in a 70-year-old man with a history of SCCA of the trachea, total laryngectomy, and tracheostomy followed by radiation 8 years before sternal resection. He underwent left lung resection for SCCA 3 years before sternal resection and had a 7-month history of tracheitis before our operation. He was operated on for sternal infection and died 11 days postoperatively of a mucous plug from mucopurulent tracheitis. Seventy percent of group 2 patients had late complications. Four patients required prolonged wound care. Two other patients required revision of their wounds at 8 and 14 months postoperatively. One patieqt had recurrent infection. Our late death occurred in a 76-year-old woman who was operated on for local recurrence 5 years after a modified radical mastectomy for invasive medullary ductal carcinoma. Her postoperative course was complicated by pneumonia, Clostridia dificile colitis, and prolonged wound care. While she was receiving postoperative adjuvant chemotherapy and radiation therapy a wound infection developed and she hemorrhaged acutely under her flap 2% months postoperatively, resulting in her death. Follow-up Half of group 1 patients are still alive, ranging from 9 to 82 months postoperatively (mean, 32.6 months). The other half died from 6 to 95 months postoperatively (mean, 39.4 months). Causes included metastatic disease in 2, stroke in 1, pulmonary embolus in 1, and congestive heart failure in 1. In group 2, 4 patients are still alive, ranging from 16 to 74 months postoperatively (mean, 35.5 months). (3f these, 2 patients had pectus excavatum and 2 had sternal infections. Six patients died, including all 5 with recurrent breast, ranging from 11 days to 15 months postoperatively (mean, 7.4 months). Comment Before undertaking sternal resections, we routinely perform pulmonary function tests. All patients received conventional chest roentgenography, which occasionally detects a defect. For evaluating congenital lesions, conventional radiography generally reveals the pathologic entity most clearly. Most benign tumors can be adequately recognized on plain films, although CT may be indicated to show the exact nature of the lesion. For the most part CT or magnetic resonance imaging is necessary to precisely define the anatomic location of lesions and help plan the extent of resection [7]. In patients with recurrent breast, CT not only helps stage recurrent disease, but occasionally shows other sites of disease not clinically suspected [8, 91. With the exception of breast and prostate, most metastatic lesions to the bony thorax are lytic. Plain films are relatively insensitive for detection of lytic lesions compared with CT scans [7]. Computed tomography can accurately localize and detect

Ann Thorac Surg 1993;5583&13 MANSOUR ET AL 841 the degree of organization of an abscess and delineate pyogenic osteomyelitis of the ribs and sternum. Adjacent lung, pleura, and mediastinal structures are also well evaluated with this modality. Computed tomography is the test of choice for evaluating patients with complications of median sternotomy. Although most of our patients were evaluated with CT scan, magnetic resonance imaging has better contrast resolution and the ability to image multiple planes, thus offering certain advantages over CT in assessment of chest wall invasive tumors [lo, 111. Infection was the most common indication for sternal resection in our series, and ranged from 0.39 to 8.4% in other series [12]. The majority of our patients had previous multiple attempts at incision and debridement with or without previous muscle flap closures of their sternal wound. Thus sternal resection was performed only after failure of less radical measures for an established infection. Six of our patients underwent sternal resections for recurrent breast. These patients tended to require more extensive resections, as evidenced by 5 of 6 requiring complete sternal resections and 4 requiring pericardiectomy. Others report 45% to 85% of patients with local recurrence would have had a previous or simultaneous distant metastasis [13-161, yet many of these patients tend to have prolonged survival [17]. Chest wall resection for local failure provides palliation for pain and removal of an ugly, ulcerated mass, thus improving quality of life. It also gives the best opportunity for local control when combined with adjuvant chemotherapy [18]. However, in our limited experience, half of these patients died within 5 months of the operation and, accordingly, we caution that careful preoperative selection should be exercised in this group. We had 5 patients with sternal tumors including a primary osteosarcoma, an eosinophilic granuloma, a tracheal recurrence of SCCA from a laryngeal primary, and 2 cases of metastatic tumor from unknown primary tumors. Four were diagnosed by either fine-needle aspiration or Tru-Cut needle biopsy. Four required partial sternal resection and 4 of 5 were reconstructed with pectoralis muscle flaps. Tumors of the sternum are frequently metastatic from breast, thyroid, or kidney [19, 201. Testicular, lung, stomach, and colon carcinomas also metastasize to the sternum [20]. Dahlin and Unni [21] reported that primary neoplasms of the manubrium and sternum constitute 15% of all primary chest wall bone tumors. The majority are chondrosarcomas, myelomas, malignant lymphomas, and osteogenic sarcomas. The prognosis for osteosarcoma of the chest wall is worse than that for chondrosarcoma [22, 231. About 60% of chest wall tumors are malignant [24]. Most primary malignant chest wall tumors are sarcomas. Osteosarcoma of the bony thorax is less common than chondrosarcoma and constitutes 6% of all primary malignant bone neoplasms. It occurs in teenagers and young adults. Treatment is wide resection of tumor and entire involved bone, rib, or sternum and adjacent soft tissues. Chemotherapy is controversial [25]. The main criterion for adequate local control of chest wall malignancy is wide excision. Sternal resection should include a portion of rib in addition to the affected portion of sternum, and it is desirable to keep a part of the sternum intact if the margin of resection safely allows it. Any attached structures such as lung, thymus, pericardium, or chest wall muscles should also be excised. Our patient with eosinophilic granuloma underwent partial sternal resection at 6 years of age. By age 8 years the process involved multiple bony sites in a malignant fashion. Ramming and associates [26] believe that age is not a significant factor in extensive chest wall operations. Eosinophilic granuloma, part of a spectrum called histiocytosis X, is only limited to bone involvement. It invades the reticuloendothelial system, consisting of a mixed inflammatory infiltrate of eosinophils and histiocytes. Excision alone results in cure for solitary lesions. However for multiple lesions, low-dose irradiation of each lesion is helpful [El. Normally pectus excavatum is surgically corrected by sternal eversion [27] or subperichondral resection with or without internal fixation [28]. However, both of our patients previously received sternal procedures, presenting us with unique challenges. The first had undergone coronary artery bypass grafting 10 years earlier and now required an aortic valve replacement. Because of fear that redo sternotomy was too dangerous to open conventionally as the sternal wires in the pectus deformity were in close proximity to the aortic arch, an entire sternectomy with bilateral pectoralis flaps and an omental flap with Prolene mesh repair was performed [29]. The second patient had a failed initial repair of pectus excavatum 30 years earlier and presented with palpitations and dyspnea on exertion. The chest wall was markedly deformed with hills and vales appearance. He underwent complete sternectomy, with pectoralis flaps and methylmethacrylate sandwich Prolene mesh repair. Eight months postoperatively the patient required removal of the methylmethacrylate mesh prosthesis due to pain and limitation of motion with excellent cosmetic result. Pectoralis major muscle flaps were often employed to close the sternal defect for primary metastatic and benign tumors plus pectus excavatum. Pectoralis major muscle and musculocutaneous flaps are frequently selected by others for closure of the sternum and anterior mediastinum due to their proximity, reliability, and versatility, and complication rates are less than for other flaps used in thoracic reconstruction [30]. Pectoralis major muscle is used for median sternotomy infections in addition to rectus abdominis muscle. We used rectus abdominis muscle flaps for recurrent breast in 4 instances where previous modified radical mastectomies had been performed, and 3 of 4 patients had previous irradiation of the chest wall. We also used it in 2 patients with osteomyelitis who previously required irradiation of the chest wall for SCCA of the trachea or larynx. In another patient we used it for osteomyelitis on a lower sternal wound with draining parasternal sinuses. In a patient with recurrent breast who had irradiation of the chest wall and extensive adenopathy surrounding the internal mam-

842 MANSOUR ET AL Ann Thorac Surg 1993;5583&13 mary artery, the tumor required a wide resection 3 cm to the right and lateral to the costal margins. Thus, a latissimus dorsi muscle flap was used as an appropriate alternative along the right chest wall to fill a large defect; however, it is not ideal for lower sternal defects because the arc of rotation makes them difficult to reach [31]. Pedicled omentum was used in three instances. In 1 patient, osteomyelitis persisted despite an initial rectus abdominis flap attempt. Because of multiple and extensive retrostemal abscesses, omentum was used to fill the gap. The second patient had undergone modified radical mastectomy plus irradiation; recurrent breast invading the mediastinum developed in her, requiring pericardiectomy necessitating pedicled omentum and a split-thickness skin graft. In the third instance an infected aortic pledget site required an aortic homograft and was covered with omentum. Omentum has superb pliability and mobility, and it can fill deep, irregular spaces. It has a rich lymphatic supply thought to aid in clearing infection. It also has an abundant low-pressure, high-flow vascular supply, which transports inflammatory cells and fibroblasts, promotes angiogenesis, and revascularizes compromised tissues, thus aiding wound healing [32]. Our various muscle and omental flaps provided soft tissue reconstruction. For stabilization of the bony thorax we prefer Prolene mesh. Prolene mesh can be sutured to the surrounding skeletal defect tightly enough to be semirigid and allows ingrowth of tissue so that the mesh becomes incorporated into the chest wall [33]. In our most recent 2 patients with sternal infection we used absorbable Vicryl mesh, thus negating a persistent foreign body in the face of potential residual infection. Half the patients in group 1 were extubated within 24 hours of operation compared with only 1 patient in group 2. The longer intubation times were thought to be necessary to allow the flail segment to firm up adequately for effective respiratory mechanics. Group 2 patients stayed twice as long in the intensive care unit compared with group 1 patients. This was due to longer intubation times and more complicated recoveries. In the partial sternectomy group, among the 3 patients with prolonged intubation times, 1 also required wedge resections of the right middle and upper lobes. A second patient with hypertension, insulin-dependent diabetes, renal failure, and peripheral vascular disease had a postoperative course complicated by lower extremity ischemia requiring a below-knee amputation. The third patient had an aortic homograft placed in addition to his omental flap. Thus, all 3 group 1 patients with longer intubation times had extenuating circumstances to explain their prolonged intubation times. References 1. Holden JS. Sarcoma of the sternum. Br Med J 1878;11:358. 2. Richardson WG. Enchondroma of the manubrium sferni successfully removed by operation. Br Med J 1913;1:9854. 3. Baue AE. Total resection of the sternum. J Thorac Cardiovasc Surg 1960;45:559-63. 4. Mansour KA, Coleman JJ. Chest reconstruction after massive chest wall resections. Emory U J Med 1989;3:132. 5. Brodin H, Linden K. Resection of the whole of the sternum and the cartilaginous parts of costae I-IV. A case report. Acta Chir Scand 1959;118:13-5. 6. Dingman RO, Argenta LC. Reconstruction of the chest wall. Ann Thorac Surg 1981;32:202-8. 7. Schaefer PS, Burton BS. Radiologic evaluation of chest-wall lesions. Surg Clin North Am 1989;69911-46. 8. Lindfors KK, Meyer JE, Busse PM, et al. CT evaluation of local and regional breast recurrence. AJR 1985;145: 833-7. 9. Villari N, Fargnoli R, Mungai R. CT evaluation of chest wall recurrence in breast. Eur J Radio1 1985;5:206-8. 10. Musset D, Grenier P, Carette MF, et al. Primary lung staging: prospective comparative study of MR imaging with CT. Radi010gy 1986;160:607-11. 11. Webb WR. MR imaging in the evaluation and staging of lung. Semin Ultrasound CT MR 1988;9:53-66. 12. Fairchild PG, Gantz NN. Mediastinal and sternal infections. Cardiac Surg State Art Rev 1988;2407. 13. Auchincloss H. The nature of local recurrence following radical mastectomy. Cancer 1958;11:611-9. 14. Marshall KA, Redfern A, Cady B. Local recurrence of carcinoma of the breast. Surg Gynecol Obstet 1974;139406-8. 15. Bruce J, Carter DC, Fraser J. Patterns of recurrent disease in breast. Lancet 1970;1:433-5. 16. Donegan WL, Perez-Mesa CM, Watson RF. A biostatistical study of locally recurrent breast carcinoma. Surg Gynecol Obstet 1966;122:52940. 17. Ryan MB, McMurtrey MJ, Roth JA. Current management of chest-wall tumors. Surg Clin North Am 1989;69:1061-80. 18. Zoetmulder FAN, von Dongen JA. Chest wall resection in the treatment of local recurrence of breast. Eur J Surg Oncol 1988;14127. 19. Boyd A. Tumors of the chest wall. In: Hood RM, Anman EM, Boyd AD, et al, eds. Surgical diseases of the pleura and chest wall. Philadelphia: Saunders, 198623940. 20. Macey HB, Phalen GS. Metastatic lesions of the sternum. Surg Gynecol Obstet 1943;76453-5. 21. Dahlin DC, Unni KK. Bone tumors: general aspects and data on 8,542 cases. Springfield: Charles C. Thomas, 1986:8-13. 22. Groff DB, Adkins PC. Chest wall tumors. Ann Thorac Surg 1967;4:2Ml. 23. Teitelbaum SL. Twenty years experience with intrinsic tumors of the bony thorax at a large institution. J Thorac Cardiovasc Surg 1972;63:776-82. 24. Pass HI. Primary and metastatic chest wall tumors. In: Roth JA, Ruckdeschel JC, Weisenburger TH, eds. Thoracic oncology. Philadelphia: Saunders, 1989:546-65. 25. Shields TW. General thoracic surgery. Philadelphia: Lea & Febiger, 1989:556. 26. Ramming KP, Holmes EC, Zarem HA, Lesavoy MA, Morton DL. Surgical management and reconstruction of extensive chest wall malignancies. Am J Surg 1982;144:14&52. 27. Wada J, Ikeda K. Clinical experience with 306 funnel chest operations. Int Surg 1972;57707-10. 28. Ravitch MM. The chest wall. In: Welch KJ, Randolph JG, Ravitch MM, et a1 eds. Pediatric surgery. Chicago: Year Book Medical, 1986563-88. 29. Gould WL, Jett GK, Bostwick J, Jones EL, Mansour KA. Simultaneous repair of severe pectus excavatum and aortic valve replacement following previous open-heart surgery. Ann Thorac Surg 1988;45:824. 30. Tobin GR. Pectoralis major muscle-musculocutaneous flap for chest-wall reconstruction. Surg Clin North Am 1989;69: 991-1027. 31. Moelleken BRW, Mathes SA, Chang N. Latissimus dorsi muscle-musculocutaneous flap in chest-wall reconstruction. Surg Clin North Am 1989;69:977-89. 32. Fix RJ, Vasconez LO. Use of the omentum in chest-wall reconstruction. Surg Clin North Am 1989;691029-46. 33. King RM, Pairolero PC, Trastek VF, et al. Primary chest wall tumors: factors affecting survival. Ann Thorac Surg 1986;41: 597-601.

Ann Thorac Surg 1993;55:83-3 MANSOUR ET AL 843 DISCUSSION DR JOE B. PUTNAM, JR (Houston, TX): I enjoyed the presentation. The experience at The University of Texas M.D. Anderson Cancer Center closely approximates the Emory experience, however fewer massive infections were treated. Eighteen patients underwent total or partial sternal resection between 1984 and 1992. Two patients had osteoradionecrosis of the sternum after radiation therapy; 16 patients were treated for primary chest wall neoplasm (n = 6), local recurrence (n = 5), or metastasis (n = 5). Histology encountered was breast (n = 3, squamous cell carcinoma (n = 3), osteogenic sarcoma (n = 3), chondrosarcoma (n = 2), and angiosarcoma (n = 1). Reconstruction was always performed for complete sternectomies, and for other defects as decided by the thoracic and plastic surgeon. Reconstruction was by rotational flaps in 7 patients, prosthetic material such as Marlex or Marlex-methylmethacrylate in 4, and nothing or not specified in 7. We have struggled with the selection of patients and the extent of operation in these patients. How did the preoperative pulmonary function influence the choice of operation? On what did you base the extent of your operation-was there a difference in patient treatment for compared with infection? Finally, has local recurrence been a significant problem? DR EDWARD R. MUNNELL (Oklahoma City, OK): In these cases of infection, were you just talking about osteomyelitis of the sternum or did these patients also had costal chondritis? In regard to total resection, Dr Anderson mentioned that you had divided the cartilages but did not completely resect them, which that leads to the second question. In the patient who had a recurrent infection, did the infection involve residual costal cartilage or not? DR MANSOUR Doctor Putnam, thank you for your discussion. We had infection after coronary bypass operations, and that is really what started the whole series, infection in the sternum and beneath the sternum. These patients actually go to the plastic surgeons first, and they had from two to four procedures before they came for the radical resection. So we realize that nothing short of total resection of the sternum will cure the problem. To answer Dr Munnell, in that regard, we resect the costal cartilages far away from the sternum because of two reasons. First, we do not want to injure the internal mammary artery, and second, we do not want to enter the heart. So we enter the chest in the third space, then we go up to the second and first space and head of the clavicle, down to the xiphoid, and then across the other side. We stay away from the midline, where there is no pericardium and the sternum is totally adherent to the heart. We start from below and dissect the sternum, shave it, so to speak, with the Bovie off the heart. In these situations, we have cardiopulmonary bypass standby and the groin prepared in the field. That is in regard to infection. In regard to the extent of resection for malignant disease, we go at least 2.5 cm beyond the margin of the tumor and the irradiated skin. The third question regarded how low or how bad pulmonary functions were before sternal resection. We usually measure pulmonary functions as a routine base study, especially in patients in whom we anticipate invasion of the underlying lung by these tumors. So to answer the question, we follow the same requirements as for any pulmonary resection. I take a forced expiratory volume in 1 second of 1 L or more for a lobe and 2 L or more for a lung, or an anticipated residual forced expiratory volume in 1 second of no less than 800 ml after resection. Finally, local recurrence after radical sternectomy has not been a significant problem in our experience. DR MUNNELL: Doctor Mansour, you did not answer my question. Have you had any cases of costal chondritis in this group? DR MANSOUR If you are talking about the infection group, we encountered some patients with costochondritis, and in these patients the costal cartilages were removed in addition to the sternal resection. However, not all patients with osteomyelitis of the sternum had associated costochondritis. DR MUNNELL Have any of them had chondritis as well as osteomyeli tis? DR MANSOUR Yes.