TUESDAY MORNING, MAY 8, 1990

7:30 a.m. FORUM SESSION - Sheraton Ballroom

F1. CO₂ Laser-Assisted Anastomoses in Internal Mammary to Coronary Artery Bypass Grafts in Growing Piglets: A Five Months Follow-Up

SEISUKE NAKATA*, PETER E. BLUNDELL,

CHARLES D. CAMPBELL and ROBERT L. REPLOGLE

Tokyo, Japan; Montreal, Quebec, Canada and Chicago, Illinois

This study was designed to evaluate the long-term complications and growth capabilities of laser-assisted internal mammary to coronary artery (IM-C) bypass anastomosis in growing piglets. Eight Yorkshire piglets weighing an average of 15 Kg. underwent IM-C bypass operation using ex-tracorporeal circulation. In one-half the animals (Group I) IM-C anastomosis was performed using a CO₂ laser (65 mW, Tohoku Ricoh, TC 3C 100) to weld tissues between 4 polypropylene stay sutures. The laser technique will be demonstrated by videotape. The remaining animals (Group II) had conventional anastomoses using 7-0 polypropylene. Animals were sacrificed when their body weights reached 60 Kg. (average 5 months) and the anastomoses were examined angiographically and microscopically for patency, diameter, internal surface and aneurysm formation. Flow through the internal mammary artery was measured using an electro-magnetic flow meter at the time of surgery and just prior to sacrifice.

Results: In both Groups I and II the five month patency in growing animals was 100%, but in Group II, 2 out of 4 anastomoses were almost occluded by intimal hyperplasia. Aneurysm formation was not seen in either group. Diameter of the anastomotic site in Group I animals grew 1.8 to 2.8 ± 0.3 mm. whereas in Group II it decreased from 1.8 to 1.4 ± 0.4 mm. (p<0.05, Group I vs II). The calibre of the normal coronary arteries increased from 1.8 to 3.0 mm. in diameter (normal growth). In Group I the flow through the internal mammary artery increased from 30 to 120 ml/min while in Group II the increase was from 30 to 70 ml/min.

Conclusions: Laser-assisted IM-C bypass anastomoses showed no long-term complications and are superior in terms of growth potential and lessened intimal hyperplasia to conventional suturing techniques. This technique may have a useful application in patients with small coronary arteries and in pediatric coronary artery surgery.

*By Invitation

F2. Pericardial Influence on Internal Defibrillation Energy Requirements

JOHN H. LEMMER, JR.*, LUKE A. FABER*,

D. JAMES MARIANO*, THOMAS A. DREWS* and

MICHAEL G. KIENZLE*

Iowa City, Iowa

Sponsored by: Douglas M. Behrendt, Iowa City, Iowa

There is no consensus regarding optimal positioning of patch leads for automatic implantable cardioverter-defibrillator (AICD) implantation. We compared energy (joules; J) required for 50% and 80% successful defibrilla-tion (E50 and E80; probit analysis) with titanium mesh patch leads (13.9 cm²) outside intact normal pericardium and directly on the epicardium in 13 open chest dogs. Talc was then instilled into the pericardial space to stimulate adhesion formation and pericardial thickening (1-5 mm). After over 3 weeks of recovery, thoracotomy and defibrillation testing were repeated with the patch leads outside the thickened adherent pericardium. Results:

There were no significant differences in defibrillation energy requirements between locations (p>0.10). In addition, comparison of electrical impedence measurements at 10J were made. No significant differences in impedence at the different locations were present (p>0.30).

We conclude that defibrillation energy requirements are not altered by the presence of normal pericardium between the patch leads and the heart as compared to being placed directly on the epicardium in this animal model. Furthermore, thickened adherent pericardium interposed between the patch leads and the heart does not increase defibrillation energy requirements. These data suggest that placement of AICD patch leads outside the pericardium in patients (including those with pericardial adhesions from previous cardiac surgery) will not adversely affect defibrillation efficacy and thus can simplify the implantation procedure.

*By Invitation

F3. Deleterious Effects of High Initial Reperfusion Pressures After Ischemia With Evidence for a Mechanism Involving Endothelial Dysfunction

KAZUO SAWATARI*, TADASHI FUJIWARA *,

TERRY A. KURTTS*, WILLIAM D. ANDERSON* and

JOHN E. MAYER, JR.

Boston, Massachusetts; Tokyo, Japan; Birmingham, Alabama

and Nashville, Tennessee

Conditions under which myocardial blood flow is restored after ischemia (Isch) have an impact on functional recovery, but the mechanisms involved in reperfusion (Rep) injury remain incompletely defined. To assess the effect of intravascular pressure (P) during Rep, isolated neonatal lamb hearts (n = 16) underwent 2 hrs of Isch with K+ cardioplegia (15°C) with 8 reperfused at low P (20mmHg X 10 min,40mmHg X 10 min, then 60mmHg) and 8 at high P (60mmHg). Control hearts (n = 16) were cooled (15°C) without Isch and rewarmed at low or high P. Recovery (Rec) of function was assessed by measuring peak developed P (PDP) and end diastolic P (EDP) during isovolumic contraction at an intra-ventricular balloon volume V'° which caused an EDP of lOmmHg before Isch. In controls (no Isch), high vs. low P during rewarming had no effect on PDP or EDP. High P post Isch resulted in 68% Rec of PDP vs. 81% with low P (p<.01). EDP at V¹⁰ was higher (p<.01) with high P (17.8mmHg) than low P (11mHg). Resting coronary blood flow (CBF) was equal in high and low P groups pre-Isch, but post-Isch CBF was 75% of baseline with high P and 128% of baseline with low P (p<.05). To explore the mechanism of the high P effect, 3 other groups (n = 24) had the same 2 hr Isch insult. Endothelial function was assessed in 2 groups by the CBF response to infusion of 10^-6M acetylcholine (Ach) which causes vasodilation, when endothelium (E) is intact, by release of E-derived relaxation factor. Vascular smooth muscle responsiveness was assessed in a third group with high Rep P by infusion of 5mcg/min nitroglycerine (TNG). Before Isch, Ach caused 25% increase in CBF in both high and low P hearts. Post-Isch, Ach caused 23% increase in CBF in low P hearts, but a 7% fall in CBF in high P hearts (p<.005). TNG during Rep resulted in 134% Rec of CBF and 80% Rec of PDP despite high Rep P (p = ns vs low P). These results show that 1) high Rep P after Isch reduces Rec of systolic and diastolic function, reduces post-Isch CBF, and reduces E-dependent coronary vasodilation, 2) Rec of mechanical function and CBF after high Rep P can be improved with the smooth muscle relaxant TNG, and 3) endothelial dysfunction with loss of regulation of coronary smooth muscle tone may be an important mechanism in the mechanical dysfunction associated with high Rep P.

*By Invitation

F4. Reduction of Infarct Size by Systemic Amino Acid Supplementation During Reperfusion

RICHARD M. ENGELMAN, JOHN A. ROUSOU,

JOSEPH E. FLACK*, JAISHIMA IYENGAR*,

YUTAKA KIMURA * and DIPAK K. DAS*

Springfield, Massachusetts and Farmington, Connecticut

The amino acids, aspartate and glutamate, in combination were evaluated as a means of reducing infarct size and improving cardiac function during reperfusion in an intact pig suffering an acute antero-septal infarct. Three groups of 6 pigs each were randomly studied, control (no amino acids), 3 mM and 13 mM aspartate/glutamate. The LAD coronary artery was occluded distal to its first diagonal branch for 60 minutes followed by 6 hour reperfusion. Aspartate/glutamate were administered systemically immediately prior to reperfusion. The following parameters were measured: infarct size and % area at risk, global metabolic function (coronary blood flow, cardiac output and 0₂ consumption), global and regional myocardial function, and tissue parameters of metabolic function (high energy phosphate, acetyl CoA and long chain acyl CoA).

The results clearly showed a significant decrease in infarct size from 61% of the area at risk in control pigs to 37% in both 3 mm and 13 mm amino acid (AA) groups. Cardiac output and coronary blood flow were not affected by the use of AA relative to the control group. Only global 0₂ consumption was significantly decreased in control pigs during reperfusion. Global LV mechanical function was not adversely affected by the infarct and was not altered by AA administration. Regional function, however, was significantly decreased by LAD occlusion in all groups to near 20% and only significantly recovered to 64% in the 13 mm AA group while remaining depressed at 22% in the control and 3 mm AA groups. High energy phosphate and acetyl CoA measurements documented significant increases in the AA groups relative to control and progressively greater levels with 13 than with 3 mm AA. Long chain acyl CoA was unchanged between the three groups.

The conclusions of this study strongly support aspartate/glutamate supplementation for stunned, reperfused myocardium. It is apparent that the effect of AA supplementation on glycolysis is directly tranlated into improved regional function and reduced infarct size.

*By Invitation

F5. Donor Hearts With Impaired Hemodynamics: Benefit of Warm Substrate Enriched Blood Cardioplegia Induction During Cardiac Harvesting

DENIS B. TIXIER*, GEORG F. MATHEIS*,

GERALD D. BUCKBERG and HELEN H. YOUNG*

Los Angeles, California

Brain dead donors frequently show circulatory deterioration, and often require so much inotropic support that the donor heart is of questionable value. This experimental study quantifies the cardiac metabolic consequences of brain death, and the role of warm blood cardioplegic induction in improving the quality of potential donor hearts with impaired hemodynamics.

Twelve dogs were subjected to brain death by interrupting cerebral blood flow (ligation of innominate artery, carotid arteries and superior vena cava) and followed for up to 6 hours. Each showed progressive hemodynamic deterioration, requiring inotropic support (Dopamine, calcium, and Epinephrine) and large amounts of volume replacement (Hespan) to suport the circulation (maintain mean arterial blood pressure > 60 mmHg). Biopsies were taken after 6 hours or when irreversible ventricular fibrillation occurred and analyzed for ATP, CP, glycogen, glutamate, and lactate. In 6 dogs, the aorta was then clamped and a 10 minute infusion of warm (37°C) substrate enriched aspartate/glutamate blood cardioplegia (using the dog's own blood) was given by roller pump to simulate warm induction during the harvesting process. Biopsies were then repeated.

Myocardial metabolism, expressed as percent of control values, during brain death was characterized by 1) moderate energy depletion: ATP fell 25 ± 8%, CP fell 55 ± 15%*,2) substrate depletion: tissue glutamate fell 48* ± 9.5%, glycogen fell 66* ± 7.5%, and 3) evidence of anaerobic metabolism: lactate increased 374* ± 95%. Warm induction of blood cardioplegia in these energy and substrate depleted ischemic hearts showed, a) return of CP to normal (113 ± 16.8%), b) replenishment of glutamate (201 ± 24% of control)*, and c) 43 ± 14% reduction in myocardial lactate content.**

These data suggest that brain dead donors requiring inotropic support sustain energy and substrate depletion and ischemic damage that can be reversed by a brief period of warm substrate enriched blood cardioplegic induction prior to harvesting. Warm substrate enriched blood cardioplegia may increase the tolerance to subsequent ischemia during organ storage, provide potential expansion of the donor pool, and/or improve the potential function of hearts harvested from hemodynamically impaired donors.

*p < 0.05 vs control: mean ± standard error of the mean

**p < 0.05 vs brain death animals

*By Invitation

F6. Reduced Reperfusion Injury by Iron Chelation in Canine Single Lung Transplantation

JOHN V. CONTE*, NEVINM. KATZ,

MARIE L. FOEGH*, JAMES ST. LOUIS*,

TINA WALSTROM*, PETER W. RAMWELL*

and ROBERT B. WALLACE

Washington, D.C.

Reperfusion injury impairs lung preservation and is caused by the generation of oxygen derived free radicals. The generation of hydroxyl radicals (- OH) via the Haber Weis reaction requires the presence of free iron. We investigated the effect of a clinically used iron chelator, deferoxamine (D), on reperfusion injury by measuring oxygen tension (pO2), pulmonary vascular resistance (PVR), alveolar arterial oxygen difference (A-a), and dynamic lung compliance (DLC) in two canine models of left single lung transplantation.

Twelve dogs received transplants following 22 hour preservation in modified Collins solution (MCS). Lungs were inflated, flushed with preservation solution and stored at 10°C with the bronchus, pulmonary artery (PA), and atrial cuff clamped. In six experiments donor and recipient received 10mg/kg D prior to harvest, prior to transplantation and in the preservation solution. Data were recorded following thoracotomy, pneumonectomy, transplantation and for six hours following ligation of non-transplanted PA. After 6 hours of ventilation with 100% 02 the mean p02 was 175.1 ± 40 mmHg vs. 71.1 ± 23 for the controls (C) (p<0.001), A-a was 502.3 ± 40 in the D group vs. 606. ± 24 mmHg for C (p<0.035). The percent change in compliance during the reperfusion period was 26.9% for the D group vs. 30.4% forC(p = NS)

Eight dogs received lungs following 8 hours of storage in MCS. Four of these dogs received deferoxamine as above. Dogs were immunosuppressed with azathioprine and cyclosporine and data was recorded after 1 hour of reperfusion and from both the native and transplanted lungs following inflation of PA occluders the day of surgery and again on post op day 5. On post op day 5 with the dogs ventilated with 100% 02 mean p02 was 254.3 ± 35 mmHg for the D group and 89.9 ± for C (p<0.04) and the A-a was 467 ± 73 dyne/sec/cm/ - 5 for C (p<0.02). Changes in compliance were similar in both groups.

This data is important because it shows that deferoxamine reduces reperfusion injury, and that the effects are lasting. Unlike other free radical scavengers, D is not a protein and not subject to the potential complications of parenteral protein administration. Additionally because it blocks the hydroxyl radical, its effects may be additive to other scavengers which block the superoxide radical.

*By Invitation

F7. Alterations of Bronchoalveolar Lymphocyte Cytolytic Functions Following Lung Allograft Rejection and Infection

HANI SHENNIB*, DAO M. NGUYEN*,

ISABELLE RAJOTTE* and DAVID S. MULDER

Montreal, Quebec, Canada

Differentiation of rejection from infection of lung allografts remains difficult. We investigated the effects of these two pathologic entities on the cytolytic activity of bronchoalveolar lavage (BAL) lymphocytes. Left lung transplantation was performed on size-matched mongrel dogs (n = 12). All animals received Cyclosporin A, Azathiopin, and Prednisone for 2 weeks. Four recipients developed left lower lobe Gram negative pneumonia within 10 days following transplantation. Remaining 8 dogs had their immunosup-pression reduced and progressed gradually to severe rejection. Cytolytic activity of left lung lymphocytes obtained by serial BAL's was assessed by the lectin dependent cell mediated cytotoxicity (LDCMC) and natural killer (NK) assays. Changes were correlated with histopathologic findings from open lung biopsy.

(Mean ± SEM % lysis at 50:1 effector to target ratio, *p<0.001 Rejection vs. Infection).

LDCMC activity progressively and drastically increased as rejection advanced. There was no detectable NK activity in rejecting allografts. Peripheral blood lymphocytes however, in contrast to BAL lymphocytes, shared an increase in NK and LDCMC activity in both settings of rejection and infection. We conclude that lung allograft rejection and infection resulted in completely different alterations of BAL lymphocyte cytolytic activity. Brochoalveolar lavage and functional studies of lung lymphocytes, in contrast to blood lymphocytes, may be a useful method in distinguishing allograft rejection from infection.

*By Invitation

F8. In Vitro, In Vivo Phototherapy of Transfected Human Lung Cancer

HARVEY I. PASS*, STEVEN EVANS*,

WILBERTA. MATTHEWS*, ROGER PERRY*,

JACK A. ROTH and PA UL SMITH*

Bethesda, Maryland and Houston, Texas

Photodynamic therapy (PDT) with dihematoporphyrin ether (P-II) sensitizes malignant cells to damage by 630 nM light. The in vitro, in vivo PDT sensitivity of a transfected human lung carcinoma line (45-342) was studied to establish a new PDT model. Using the colony formation assay neither light alone nor P-II alone affected 45-342 cell survival. Energy-dependent PDT effects were seen in vitro in P-II-incubated light exposed cells (90% cytotox-icity = 950 J/m², 99% cytotoxicity = 1575 J/m², p₂ < .05), and cellular [P-II] increased to a maximum of 2.77 ug/10⁶ cells at 4 hours.

Ideal route (IV vs IP) of P-II administration (10 mg/kg), tissue distribution kinetics, and PDT effects were investigated in 133 nu/nu mice allografted subcutaneously with 45-342. IV P-II was superior to IP P-II 24 hours after injection (32.4 ± 2.1 ug P-II/g dry wt vs 17.7 ± 1.9 ug P-II/g dry wt, p₂ < .01), and was more precise (coefficient of variation p₂ < .035). Tumor retained high P-II levels (ug/dry wt), as seen below.

(*p₂<.05 from skin and muscle, #p₂<.001 from tumor at 48 hours)

Tumor/tissue ratios were significantly elevated at 24 hours (tumor/skin = 3.0, tumor/muscle = 6.23) compared to 48 hours (tumor/skin = 1.8, tumor/muscle = 3.8) [p₂<.005]. PDT (.3W/cm², 150J/cm²) 24 hours after IV P-II administration caused progressive coagulative tumor necrosis (n = 16) and tumor regression.

These data establish the PDT sensitivity of transfected human lung cancer in vitro and in vivo. Consistent levels of tissue P-II can be determined after IV administration, and preferential tumor P-II retention observed. This tumor model may be useful to study in vitro mechanisms of PDT action and in vivo PDT effects to insure maximum selective tumor kill.

*By Invitation

F9. Cardiac 5'Nucleotidase Increases With Age and Inversely Correlates With Recovery from Ischemia

MICHAEL A. GROSSO*, ANIRBAN BANERJEE*,

JOHN A. ST. CYR*, JAMES M. BROWN*,

GLENN J. WHITMAN*, DAVID N. CAMPBELL*,

DAVID R. CLARKE and ALDEN H. HARKEN

Denver, Colorado

The metabolic/biochemical basis for the enhanced tolerance of immature hearts to ischemia remains to be elucidated. Documented loss of high energy phosphate nucleotides occurs during ischemia/reperfusion in mature hearts through the breakdown of ATP, ADP and AMP (non-diffusable) to Adenosine (Ado) (freely diffusable). The enzyme responsible for the conversion of AMP to Ado is 5'-Nucleotidase (5'NT). Previous work has shown this nucleotide loss to be blunted in ischemic immature hearts. We therefore hyothesized lower levels of this enzyme in immature versus mature myocardium. The purposes of this study were (1) to document 5'NT activities in immature and mature rabbit myocardium and (2) correlate differences of 5'NT with functional recovery from ischemia. Neonatal (5-7 days) and adult (4-6 months) rabbit hearts were isolated and perfused (retrograde Langendorff). A left ventricular balloon measured functional parameters. Hearts were subjected to 10 min equilibration followed by freeze-clamping (5'NT assay) or 20 min global 37° ischemia and reperfusion. Assay for

Functional recovery from ischemia inversely correlated with age (r = 0.99) and 5'NT(r = 0.96)

We conclude: (1) 5'Nucleotidase levels are markedly diminished in neonatal versus adult myocardium and (2) functional recovery following ischemia inversely correlates with age and 5'-nucleotidase activity.

*By Invitation

F10. The Mechanism of Heart Failure Caused by Cardiac Allograft Rejection

VERDI J. DiSESA *, PAOLO MASETTI*,

JAMES MARSH*, FREDERICK J. SHOEN* and

LAWRENCE H. COHN

Boston, Massachusetts

Rejection of the cardiac allograft may cause reversible myocadial failure by a poorly defined mechanism. To study this phenomenon, heterotopic cardiac transplantation was performed in the Lewis rat using Lewis X Brown Norway (allografts) or Lewis (isografts) donors. Without immunosuppres-sion, allografts are rejected in 6-8 days. At 3 days cardiac grafts were ex-planted and underwent functional measurements (Langendorff apparatus) or pathologic and biochemical examination. Three-day isografts (n = 9) had minimal histologic changes. Three-day allografts (n = 13) showed rejection with lymphocytic infiltrate and myocyte necrosis. Functional studies (n = 11) included heart rate (HR) (beats/min), cardiac output (CO) (ml/min), coronary flow (CF) (ml/min), and stroke work (SW) (dynes/cm) at baseline and in response to isoproterenol (3 x 10^-8 M/ml):

Allograft ATP levels (n = 4) were normal (mean 41.9 nmol/mg).

Functional, biochemical, and pathological studies in this small animal model demonstrate diminished contractile reserve despite normal baseline function, chronotropic response, coronary flow and high energy stores in allografts with early but significant rejection. These data suggest that alteration in intra-cellular signals important in excitation-contraction coupling are the mechanism of heart failure caused by allograft rejection.

*By Invitation

9:00 a.m. SCIENTIFIC SESSION - Sheraton Ballroom

15. The Role of Extrapleural Pneumonectomy in Malignant Pleural Mesothelioma

VALERIE W. RUSCH*, STEVEN PIANTODOSI*,

E. CARMACK HOLMES and THE LUNG CANCER

STUDY GROUP

New York, New York; Baltimore, Maryland and Los Angeles, California

Malignant pleural mesothelioma (MPM) is usually a fatal cancer for which surgery has been the mainstay of treatment because chemotherapy and radiation are relatively ineffective in this disease. The choice of operation for MPM remains controversial. Extrapleural pneumonectomy (EPP) has been advocated because it allows complete removal of gross tumor and can be associated with long-term survival.

To evaluate EPP, we carried out a prospective multi-institutional trial in patients with biopsy-proven previously untreated MPM. Criteria for EPP were: (1) potentially completely resectable unilateral disease by CT scan (2) predicted post-resection FEV1 > IL/sec (3) no major medical problems. Patients who were not candidates for EPP underwent pleurectomy/decortica-tion (P/D) or had non-surgical treatment (NST).

Results: from 9/85 to 6/88, 83 patients (64 males, 19 females) were entered.

In univariate analyses, epithelial versus sarcomatoid and mixed histologies had a better overall survival (p = 0.02); and performance status (Karnofsky < 9) was predictive of recurrence (p = 0.02). In multivariate analysis, histology, sex, age, EPP, and performance status all had no significant impact on survival. EPP was associated with a higher incidence of relapse in distant sites (7/20, 35%), than were P/D (3/26, 11.5%) and NST (2/37, 5.4%).

Conclusions: (1) only a small proportion of all MPM patients are candidates for EPP; (2) EPP carries a significant operative mortality and does not improve overall survival compared to more conservative forms of treatment; (3) EPP alters the patterns of relapse; (4) factors previously thought to have an impact on survival in other series did not affect outcome in this trial.

*By Invitation

16. Extrapleural Pneumonectomy, CAP Chemotherapy and Radiotherapy in the Treatment of Diffuse Malignant Pleural Mesothelioma

DAVID J. SUGARBAKER*, THOMAS LEE*,

KAREN ANTMAN*, GREGOR Y COUPER *,

STEVEN MENTZER*, JOSEPHM. CORSON*,

JOHN J. COLLINS, JR., RICHARD SHEMIN

and ROBER T PUGATCH*

Boston, Massachusetts

Diffuse malignant mesothelioma has been deemed a non-operative disease in many centers. Extrapleural pneumonectomy has proven disappointing as have single and multiple drug chemotherapy with or without radiotherapy in the treatment of diffuse malignant mesothelioma. Median survivals of 10 to 15 months have been reported with two year survival rates of 10-35%.

From 1980 to 1989, 30 patients with pre-operative Stage I pleural mesothelioma (tumor confined within the capsule of the parietal pleural, i.e. involving only the ipsilateral lung, pericardium, and diaphragm) underwent extrapleural pneumonectomy with resection of pericardium and diaphragm followed by three cycles of standard CAP chemotherapy and 4500 rads of radiotherapy to the ipsilateral chest. All patients had the pathologic diagnosis reviewed prior to treatment. The age of patients was 52.4 +/- 8.0 years (range 41-68 years); 25 were men and five women.

There were two post operative deaths, (6%). One death was due to myocardial infarction defined at autopsy and the other death was clinically attributed to pulmonary embolus.

For the 28 patients surviving the operation, survival rates were 78% at one year, 52%at two years, and 32% at three years. Sixteen patients are currently alive following this trimodality therapy (median follow up 15.6 months) with one patient alive at 56 months post treatment. The 12 patients who died of disease had a median survival of 13.9 +/- 11 months.

These data suggest that radical extrapleural pneumonectomy can be performed with an acceptable operative mortality. When combined with post operative CAP chemotherapy and radiotherapy, it may provide an effective therapeutic and palliative alternative in the primary treatment of diffuse malignant mesothelioma.

*By Invitation

17. Carinal Resection for Lung Cancer

DOUGLAS J. MATHISEN and HERMES C. GRILLO

Boston, Massachusetts

The revised staging system for lung cancer classifies tumors within 2 centimeters of the carina as STAGE III B. This has implied inoperability and incurability based on technical limitations rather than advanced disease. Advances in bronchoplastic technique now allow resection of the carina with lower morbidity and mortality and possible cure. Consideration should be given to include this select group of patients in STAGE III A.

We have treated 30 patients with cancer of the lung requiring carinal resection. There were 24 males and 6 females. Histology was squamous cancer in 22 and adenocarcinoma in 7 and adenosquamous carcinoma in 1. Right thoracotomy was used in all patients. Right carinal pneumonectomy was performed in 15, resection of the carina plus the right upper lobe with preservation of the right middle and lower lobes in 7, resection of the carina alone in 6, and resection of the carina plus stump of bronchus following prior pneumonectomy in 2 all with appropriate reconstructions. Eleven patients had positive nodes. The majority of patients had postoperative radiation therapy. There were 4 postoperative deaths (anastomotic complications-1, pneumonia-2, ARDS-1). Two patients died 3 months postoperatively from anastomotic complications. Excluding the operative deaths, major complications included anastomotic stenosis requiring completion pneumonectomy-1, anastomotic leak requiring omental wrap-1, pneumonia-4, hoarseness-2. Of the operative survivors, there are 4 absolute 5 year survivors, 5 patients are alive between 3 and 5 years, 12 alive between 1 and 3 years and 3 dead from disease 1 to 3½ years.

In selected patients, carinal resection can be accomplished. The prospect for long-term survival now appreciably exceeds the operative mortality and justifies this aggressive approach.

INTERMISSION - VISIT EXHIBITS

*By Invitation

10:45 a.m. SCIENTIFIC SESSIONS

18. Esophageal Ultrasound and the Preoperative Staging of Carcinoma of the Esophagus

THOMAS W. RICE*, MICHAEL V. SIVAK*

and GREG A. BOYCE*

Cleveland, Ohio

Sponsored by: Floyd D. Loop, Cleveland, Ohio

Esophageal ultrasound [EUS] allows the esophagus and paraesophageal tissue to be viewed as five discrete layers. Lymph nodes are easily identified and their size, shape, margin, and internal structure can be assessed. This provides an alternative method of preoperative evaluation of the primary tumor status [T] and the regional lymph node status [N] of patients with carcinoma of the esophagus.

EUS was attempted in the preoperative staging of 28 patients with carcinoma of the esophagus. Six patients were not assessed because of the inability to pass the esophageal ultrasound probe through the malignant stricture. The AJCC staging system for carcinoma of the esophagus was used.

Twenty-two patients had the true T determined by pathologic review of the resected esophagus. EUS correctly identified T in 13 patients (59% accuracy). Four patients were overstaged by esophageal ultrasound; all these patients had early T, tumors confined to the submucosa. Five patients were understaged by EUS; all of these patients had advanced tumors (four T₃ and one T₄) that invaded outside the esophageal wall. Seven of the nine incorrect EUS determinations were called T₂ (three T,, three T₃, one T₄) suggesting that the borders of the muscularis propria may require careful attention when evaluated by EUS.

Twenty patients had the true N determined by pathologic review of the resected specimen. EUS correctly identified N in 14 patients (70% accuracy). Three patients were falsely identified as N, and three patients were falsely identified as N₀. The sensitivity, specificity, positive predictive value, and negative predictive value for EUS and N assessment were 70%.

EUS provides an alternative method of visualization of the esophageal wall and regional lymph nodes. Our early experience shows promise for EUS in the preoperative staging of carcinoma of the esophagus.

*By Invitation

19. An Endothoracic Endoesophageal Pull-Through Operation for Cancer of the Esophagus: Technique and Results in 68 Patients

FARROKH SAIDI*

Tehran, Iran

Sponsored by: Clement A. Hiebert, Portland, Maine

Transhiatal esophagectomy for cancer of the distal or proximal esophagus is simplified if the uninvolved muscular layer of the esophagus is left in situ. Mucosa of the normal segment is merely stripped and removed as part of the specimen. A distensible muscle tunnel remains and through it mobilized stomach or colon can be drawn to the neck. The approach is based on three anatomic features of the normal esophagus: The mucosal layer is tough, its attachment to the underlying muscle is flimsy, and vascular connections are small.

The endothoracic endoesophageal pull-through operation has been used in the palliative treatment of 42 consecutive unselected patients with cancer of the lower esophagus or cardia and three patients with cancer of the cervical esophagus. As with conventional transhiatal esophagectomy, the chest is not opened. Following standard mobilization of the stomach and resection of the tumor and regional lymph nodes, the upper and lower thoracic ends of the normal esophagus are circumcised and the mucosal layer cored out by finger dissection. Blood loss is scant.

In 28 patients the stomach, and in fourteen patients an isoperistaltic loop of colon was used for reconstruction. The overall mortality was 13%, cervical anastomotic leakage rate 13%, and survival rates at two and three years, respectively, 18% and 13%.

In a separate group of 26 patients, who required thoracotomy for middle 3rd cancer, mucosal stripping was started 10 cm above the proximal tumor edge; the mobilized stomach was then passed through this shorter muscular tunnel to allow a cervical anastomosis.

11:15 a.m. ADDRESS BY HONORED SPEAKER

"PIONEERS AND MILESTONES IN THORACIC SURGERY"

A.P. Naef, Fully, Switzerland

12:00 p.m. ADJOURN FOR LUNCH - VISIT EXHIBITS

12:00 p.m. CARDIOTHORACIC RESIDENTS' LUNCHEON

Civic Ballroom (2nd floor)