Does the Timing of Esophagectomy After Chemoradiation Affect Outcome?

  • Published on

  • View

  • Download



    loc, yourna

    D agC tcJa A.Re DavD D, JEd MD

    De tinalOn Canc

    Background. After neoadjuvant chemoradiation (CXRT)foromsuoththetheleaop


    Results. One hundred fifty patients were resected within8 wcomtivcomlogsimvadecomC








    ERALTHORACICuccessful treatment of locally advanced esophagealcancer remains challenging. Randomized trials have

    own that the addition of neoadjuvant chemoradiationXRT) leads to a survival benefit compared with surgeryne for the treatment of locally advanced esophagealcer [13]. Most of the clinical trials evaluating neoadju-

    nt therapy have stipulated that surgery be performedthin a defined time frame after completion of therapy.ditionally, this has been within 3 to 8 weeks afterpleting radiation therapy [27]. An extrapolation from

    s tradition is that surgery is recommended within adefined duration whether patients are on or off clinical

    protocol. However, for a variety of reasons, many patientsdo not undergo surgery within that time frame. Somepatients have suffered adverse events frommedical therapyor simply have not adequately recovered from their neoad-juvant therapy. Other patients have their surgery delayedfor personal or logistic reasons.Radiation-induced tumor necrosis may increase over

    time and studies of neoadjuvant radiotherapy for rectalcancer suggest that a longer interval between radiation andsurgery may actually result in improved pathologic com-plete response (pCR) and decreased postoperative morbid-ity [8, 9]. On the other hand, there are theoretical concernsthat waiting longer could make the dissection more difficultdue to increased radiation fibrosis [10]. There is also apossibility that delaying surgery could allow for tumorregrowth, increasing the risk of recurrence. The optimaltiming of surgery after CXRT for esophageal cancer isunknown. We thus decided to analyze patients who under-

    epted for publication May 3, 2011.

    sented at the Forty-seventh Annual Meeting of The Society of Thoracicgeons, San Diego, CA, Jan 31Feb 2, 2011.

    dress correspondence to Dr Kim, Department of Surgery, City of Hopecer Center, 1500 E Duarte Rd, MOB Ste 2001, Duarte, CA 91010;ail:

    2012 by The Society of Thoracic Surgeons 0003-4975/$36.00blished by Elsevier Inc doi:10.1016/j.athoracsur.2011.05.021esophageal cancer, surgery has traditionally been rec-mended to be performed within 8 weeks. However,rgery is often delayed for various reasons. Data fromer cancers suggest that delaying surgery may increasepathologic complete response rate. However, there areoretical concerns that waiting longer after radiation mayd to amore difficult operation andmore complications. Thetimal timing of esophagectomy after CXRT is unknown.Methods. From a prospective database, we analyzed6 patients with resected esophageal cancer who wereated with neoadjuvant CXRT from 2002 to 2008. Sal-ge resections were excluded from this analysis. Wepared patients who had surgery within 8 weeks ofRT and those who had surgery after 8 weeks. We usedltivariable analysis to determine whether increasederval between chemoradiation and surgery was inde-ndently associated with perioperative complication,thologic response, or overall survival.eeks and 116 were resected greater than 8 weeks afterpleting CXRT. Mean length of operation, intraopera-

    e blood loss, anastomotic leak rate, and perioperativeplication rate were similar for the two groups. Patho-ic complete response rate and overall survival were alsoilar for the two groups (p not significant). In multi-

    riable analysis, timing of surgery was not an indepen-nt predictor of perioperative complication, pathologicplete response, or overall survival.onclusions. The timing of esophagectomy after neo-

    juvant CXRT is not associated with perioperative com-cation, pathologic response, or overall survival. It mayreasonable to delay esophagectomy beyond 8 weekspatients who have not yet recovered from chemora-tion.

    (Ann Thorac Surg 2012;93:20713) 2012 by The Society of Thoracic SurgeonsGENERAL THORACIC SURGERY:The Annals of Thoracic Surgery CME Program isTo take the CME activity related to this articleindividual non-member subscription to the jou

    oes the Timing of Esophhemoradiation Affect Oue Y. Kim, MD, Arlene M. Correa, PhD, Araza J. Mehran, MD, Garrett L. Walsh, MD,ipen M. Maru, MD, Manoop S. Bhutani, Mith M. Marom, MD, Stephen G. Swisher,

    partments of Thoracic and Cardiovascular Surgery, Gastrointescology, and Radiology, The University of Texas MD Andersonated online at have either an STS member or an


    ectomy Afterome?Vaporciyan, MD, Jack A. Roth, MD,id C. Rice, MD, Jaffer A. Ajani, MD,ames Welsh, MD,, and Wayne L. Hofstetter, MD

    Medical Oncology, Pathology, Gastroenterology, Radiationer Center, Houston, Texas

  • wetutpelon











    208 KIM ET AL Ann Thorac SurgTIMING OF ESOPHAGECTOMY 2012;93:20713


    ERALTHORACICnt esophagectomy after neoadjuvant CXRT at our insti-ion to explore what effect the timing of surgery had onrioperative complications, pathologic response, andg-term outcomes.

    tients and Methods

    e retrospectively analyzed a prospectively maintainedtabase of patients with esophageal cancer who wereated with neoadjuvant CXRT followed by surgery at ourtitution from 2002 to 2008. This study was approved byr Institutional Review Board. Individual consent wasived due to the retrospective nature of the study. Patientsth either squamous cell or adenocarcinoma histologyre included. We excluded all emergency cases, redophagectomies, patients with distant metastatic disease,d patients with recent metachronous or synchronousmary cancers. Salvage cases were also excluded. Casesre designated as salvage if they underwent plannedfinitive chemoradiation. Those patients who receivedoadjuvant chemoradiation, had a complete clinical re-onse, and chose to undergo surveillance rather thanrgery were also considered salvage if they went on tove surgery after evidence of recurrence. Those patientso had planned neoadjuvant therapy and had persistentease on restagingwerenot considered salvage regardless oftime interval between chemoradiation and surgery.

    Patients were divided into 2 groups: those who hadrgery within 8 weeks of completing CXRT and thoseo had surgery greater than 8 weeks after completingRT. We compared the rates of major perioperativemplications, pathologic response, and overall survivalthe two groups. Perioperative death was defined as

    ath occurring during initial hospitalization regardlesslength or within 30 days after esophagectomy. We alsoamined the estimated blood loss, transfusion require-nt, and length of operation as surrogate markers fordifficulty of the operation. We then performed a

    ltivariable analysis to determine which factors wereependently associated with perioperative complica-n, pathologic response, and overall survival. Thoseriables with a significance of less than 0.25 on univari-le analysis were included in the regression. For theltivariable analysis, timing of surgery was examinedth as a continuous variable and a dichotomous variabletients with surgery within 8 weeks of completingRT and patients with surgery 8 weeks after CXRT).Overall, 100% of patients had endoscopic ultrasound and% had positron emission tomography (PET). Stagings according to the 6th edition American Joint CommitteeCancer guidelines. Patients who were stage IVa wereluded in the study. Platinum-based chemotherapy wasen to 69% of patients. Nearly all (98%) patients re-ved 45 Gy or a higher dose of radiation.After completing chemoradiotherapy, all patients weretaged with PET-computed tomography or computedography, and repeat endoscopy was also performed.

    tients were then reassessed by the surgeon. Surgical

    hnique was determined by tumor location and surgeonference. FigAll resected specimens underwent routine histopathol-y. If no tumor cells were identified in the specimen,luding lymph nodes, patients were classified as hav-a pCR. All other patients were classified according toAmerican Joint Committee on Cancer guidelines.

    tients were followed periodically for at least 5 years ortil death. Median potential follow-up was 55 monthsto 99 months).

    bgroup Analysise found that thereweremore patients with squamous celltology in the longer interval group. In order to eliminates potential selection bias, we performed a subgroupalysis using only patients with adenocarcinoma histol-y. In a separate subgroup analysis, we explored whethertients who had surgery greater than 12 weeks afterpletion of CXRT had different outcomes than patientso had surgery within 8 weeks of CXRT.

    tistical Analysisrvival probability analyses were performed using theplan-Meier method and were calculated from the datediagnosis to the date of death or most recent follow-up.tistical significance was assessed by the log-rank testMann-Whitney test. Univariate analyses were per-med by 2 analysis. Factors independently associatedth perioperative complication and long-term survivalre determined by logistic regression or Cox regressionalysis. The 2-tailed probability values (p values) of 0.05less were considered significant. Data were analyzedthe SPSS statistical software (SPSS, Chicago, IL).


    m 2002 to 2008, 570 esophagectomies were performedeither squamous cell or adenocarcinoma at our insti-ion. A total of 382 patients received neoadjuvantemoradiation. Two hundred sixty-six patients who metinclusion criteria were identified. One hundred fifty%) had surgery within 8 weeks and 116 (43%) hadrgery greater than 8 weeks after completing chemora-therapy (Fig 1). Among patients in the short interval1. Distribution of timing of surgery after chemoradiation (CXRT).

  • grotheme32









    The , smoche


    209Ann Thorac Surg KIM ET AL2012;93:20713 TIMING OF ESOPHAGECTOMY


    ERALTHORACICup, the median interval between completion of radio-rapy and surgery was 46 days (21 to 56 days). Thedian interval in the delayed group was 87 days (57 to2 days) (p 0.001).The majority of patients (76% in the short intervalup and 65% in the delayed group) underwent anr-Lewis approach with a separate laparotomy andracotomy with intrathoracic anastomosis, as shown inble 1. The remaining patients were split relativelyenly among three-field (9%), transhiatal (10%), andnimally invasive (11%) approaches. Four patients inch group (3% of total) underwent primary jejunalerposition and were classified as three-field.Patients in the delayed group had a higher mean agevs 57 years), were more likely to have squamous celltology (13% vs 3%), were less likely to be overweight% vs 74%), and had a higher incidence of coronaryery disease (17% vs 7%), as shown in Table 1. Theyre also significantly less likely to be treated as part of

    ble 1. Baseline Characteristics of Patients With an 8-Week orith a Greater Than 8-Week Interval Between Radiation and S

    riableInterval 8 W

    (n 150

    an age 57le 135 (90%)ucasian 132 (88%)I 25 112 (75%)I 18.524.9 38 (25%)I 18.5 0enocarcinoma 145 (97%)wer third/GE junction 145 (97%)brod performance status 2 4 (3%)brod 1 81 (54%)brod 0 65 (43%)ronary artery disease 11 (7%)PD 6 (4%)ight loss within 3 months 72 (48%)betes 21 (14%)protocol 76 (51%)

    diation dose 45 Gy 150 (100%pe of operationvor-Lewis 114 (76%)ranshiatal 13 (9%)hree-field 10 (7%)inimally invasive 13 (9%)

    mor gradeell and moderately

    ferentiated45 (34%)

    oor and undifferentiated 87 (66%)nical stageI 58 (39%)II 88 (59%)Va 4 (3%)

    re was no significant difference between the groups for alcohol use

    motherapy received, and pulmonary function test (p 0.05).

    I body mass index; COPD chronic obstructive pulmonary disease;esearch protocol (24% vs 51%). Otherwise, there werestatistically significant differences in the baseline

    aracteristics of the two groups of patients.The operative outcomes for the two groups wereilar in terms of average length of operation, thember of lymph nodes removed during the operation,d the mean estimated blood loss, although patientsthe delayed group did have a higher rate of intra-erative transfusion (22% vs 13%, p 0.05), as shownTable 2. There were no statistically significant dif-ences in postoperative mortality, pulmonary com-cation, anastomotic leak, or median length of stay.e overall complication rate was 39%. The anasto-tic leak rate was 11% in the short interval group and% in the delayed group (p 0.28). In univariablealysis, increased time to surgery, histology, Ameri-n Society of Anesthesiologists classification, andent weight loss were associated with perioperativemplication. In multivariable analysis, however, only

    s Interval Between Radiation and Surgery and Patientsry

    s Interval 8 Weeks(n 116) p Value

    60 0.0397 (84%) 0.14103 (89%) 0.8562 (53%) 0.00146 (49%)8 (7%)

    101 (87%) 0.003109 (94%) 0.384 (4%) 0.1373 (63%)39 (34%)20 (17%) 0.023 (3%) 0.7461 (53%) 0.5414 (12%) 0.7228 (24%) 0.001114 (98%) 0.19

    76 (66%) 0.2814 (12%)9 (8%)17 (15%)

    0.4940 (39%)

    62 (61%)0.34

    40 (35%)69 (60%)7 (6%)

    king history, American Society of Anesthesiologists class, type ofa rnoch



    eek)GE gastroesophageal.

  • reccoT








    PaPSSSS 1%)

    PoM 2%)P 31%)A 11%)M 1M 077

    Me l exp







    210 KIM ET AL Ann Thorac SurgTIMING OF ESOPHAGECTOMY 2012;93:20713


    ERALTHORACICent weight loss was found to be associated withmplication (p 0.02) as shown in Table 3.he shorter interval group and the delayed groupd similar rates of pCR (21% vs 22%) with a similartribution of pathologic stage (Table 2). The multi-riable analysis also failed to demonstrate an associ-on between timing of surgery and pathologic re-onse (Table 3).verall survival was similar for the 2 groups (p 0.23)shown in Figure 2. Median survival for the shorterval group was 53 months (95% confidence interval,to 77 months) and 39 months for the delayed group% confidence interval, 28 to 50 months). Likewise,re was no significant difference in disease free survivalthe two groups as shown in Figure 3. Onmultivariablealysis, age, pathologic stage, and number of positive

    ble 2. Postoperative Outcomes of Patients With an 8-Week orith a Greater Than 8-Week Interval Between Radiation and S



    eration characteristics:ntraoperative transfusion 19 (ean estimated blood loss (mL) 502ean surgery time (minutes) 390ean number of lymph nodes removed 21

    thologic stage:athologic complete response 32 (tage I 19 (tage II 65 (tage III 33 (tage IV 1 (stoperative complications:ortality 3 (ulmonary complication 47 (nastomotic leak 17 (edian length of stay (days) 1edian survival (months) 53 [3

    an values expressed with standard error of the mean. Median surviva

    ble 3. Multivariable Analysis of Independent Predictors of Corvival

    dpoint Independent Predictor

    mplication Weight lossTime to surgery (days)

    R Female genderBMI 25Squamous cell histologyTime to surgery (days)

    erall survival AgeNumber of lymph nodes positivePathologic stage IVTime to surgery (days)I body mass index; CI confidence interval; HR hazard ratioplication and pCR endpoints; pCR pathologic complete nodes were significantly associated with survival 0.05) as shown in Table 3.

    bgroup Analysesa subgroup analysis of patients with only adenocarci-ma histology, those patients who had surgery 8 weeksless after CXRT and those who had surgery greatern 8 weeks after CXRT had similar rates of anastomotick (11% vs 15%, p 0.44), perioperative complication% vs 41%, p 0.42), and pCR (20% vs 19%, p 0.76).dian survival was greater for those in the short inter-l group (53 vs 36 months), but this difference was nottistically significant (p 0.21).In an effort to evaluate whether patients who werenificantly delayed from going to surgery had differenttcomes from those undergoing standard recovery pe-

    s Interval Between Radiation and Surgery and Patientsry

    eeks Interval 8 Weeks(n 116) p Value

    26 (22%) 0.05478 29 0.44398 10 0.3720 0.9 0.21

    0.7926 (22%)10 (9%)51 (44%)27 (23%)2 (2%)

    4 (3%) 0.7041 (35%) 0.5119 (16%) 0.28

    11 0.41] 39 [2850] 0.23

    ressed with 95% confidence interval.

    cation, Pathologic Complete Response, and Overall

    OR/HR 95% CI p Value

    1.84 1.103.06 0.021.01 0.9991.015 0.072.51 1.095.82 0.032.69 1.245.84 0.014.87 1.6514.34 0.0041.002 0.991.01 0.671.026 1.0081.045 0.0051.115 1.0491.185 0.0015.105 1.50417.32 0.0091.001 0.9961.005 0.79(35Mevasta



    8 W150)

    13%) 53 8 0.8

    21%)13%)43%)22%), used for overall survival endpoint; OR odds ratio, used for

  • rioforCXsuinco24%shthacalmomowa





    inccha d21chtimlonsustucanbetheratsuesoovpetheindif



    211Ann Thorac Surg KIM ET AL2012;93:20713 TIMING OF ESOPHAGECTOMY


    ERALTHORACICds after CXRT, another subgroup analysis was per-med. Patients who had surgery 8 weeks or less afterRT were compared with the 41 patients who hadrgery greater than 12 weeks after CXRT. The patientsthe two groups had similar rates of perioperativemplication (36% vs 42%, p 0.59) and pCR (21% vs, p 0.55). The anastomotic leak rate was 11% in the

    ort interval group compared with 17% in the greatern 12-week group, but this difference was not statisti-ly significant (p 0.42). The median survival was 53nths in the short interval group compared with 46nths in the greater than 12-week group, but this alsos not statistically significant (p 0.33).


    this study, a longer interval between neoadjuvantRT and surgery was not associated with a difference instoperative morbidity, pathologic response, or overallrvival. By the criteria we measured (length of opera-n, estimated blood loss, and number of lymph nodesrieved), a longer interval did not appear to affect thenduct of the operation. The longer interval group didve a higher incidence of intraoperative transfusion;wever, this may only reflect the older age or increasedidence of coronary artery disease in this group oftients.Studies on neoadjuvant chemoradiation for rectal can-have shown conflicting data about the impact of theerval between chemoradiation and surgery. Whereasme studies have shown no differences in postoperativemplications, others have shown increased rates ofastomotic leak with shorter intervals [8, 9, 1116]. Inntrast, in our study the anastomotic leak rate wasually higher in the delayed group (16% vs 11%), buts difference was not statistically significant. The leake of 11% in the short interval group is comparable with

    2. Kaplan-Meier curve for overall survival; p studies of esophagectomy after neoadjuvantemoradiation [2, 7]. There was a trend toward in- Figased anastomotic leaks (16% vs 11%), mortality (3% vs), and pulmonary complications (35% vs 31%) in thelayed group. This may reflect the higher baseline riskcreased age, weight loss, etc) in this cohort. It is alsossible that our study lacked statistical power to confirmse small differences. The median length of stay (arker for serious complication) was similar for the twoups.

    Some data from rectal cancer have also shown anreased rate of pCR, with a longer interval betweenemoradiation and surgery. In contrast, we did not findifference in the rate of pCR. The overall pCR rate of% is comparable with other studies of neoadjuvantemoradiation and surgery [17]. We did not find that theing of surgery was associated with a difference ing-term survival or disease-free survival. The medianrvival of 46 months compares favorably with otherdies of neoadjuvant chemoradiation for esophagealcer [17]. In a similar study examining the intervaltween CXRT and surgery for squamous cell cancer ofesophagus, Ruol and colleagues [18] found that pCR

    e did not change with increased interval betweenrgery and CXRT for squamous cell carcinoma of thephagus; however, the delayed group had a lowererall recurrence rate. There was a significantly higherrcentage of patients with squamous cell carcinoma indelayed group, but eliminating the bias of histologythe subgroup analysis did not reveal any significantferences in outcome.In summary, our results do not support a longererval after neoadjuvant chemoradiation for esopha-al cancer in order to increase pathologic response. Onother hand, we also did not show any increase in

    erative morbidity for patients with longer intervals;s, it may be reasonable to delay surgery for patientso have not yet recovered from chemoradiation. Theseults should be interpreted with caution because wely included patients who underwent resection. Thisresents only a selected group of patients who received3. Kaplan-Meier curve for disease-free survival; p 0.17.

  • neultdiscermaressuexmoou

    biadidtherefore, a delay in surgery could represent a surrogateenbe

    takfrosuweopoua lthr










    An interval 7 weeks between neoadjuvant therapy andsurgery improves pathologic complete response and dis-

















    212 KIM ET AL Ann Thorac SurgTIMING OF ESOPHAGECTOMY 2012;93:20713


    ERALTHORACICdpoint. In addition, small differences in survival woulddifficult to detect with this relatively small sample size.In conclusion, we recommend that surgery be under-en at the earliest opportunity after adequate recoverym neoadjuvant therapy. But if it is necessary, our dataggest that an additional delay of surgery beyond 8eks is not associated with significantly increased peri-erative morbidity or mortality. Furthermore, althoughr data suggest that there is not an increase in pCR withonger interval to surgery, validation of these findingsough a prospective trial may be warranted.


    Gebski V, Burmeister B, Smithers BM, et al. Survival benefitsfrom neoadjuvant chemoradiotherapy or chemotherapy inoesophageal carcinoma: a meta-analysis. Lancet Oncol 2007;8:22634.Tepper J, Krasna MJ, Niedzwiecki D, et al. Phase III trial oftrimodality therapy with cisplatin, fluorouracil, radiother-apy, and surgery compared with surgery alone for esopha-geal cancer: CALGB 9781. J Clin Oncol 2008;26:108692.Walsh TN, Noonan N, Hollywood D, Kelly A, Keeling N,Hennessy TP. A comparison of multimodal therapy andsurgery for esophageal adenocarcinoma. N Engl J Med1996;335:4627.Bosset JF, Gignoux M, Triboulet JP, et al. Chemoradiother-apy followed by surgery compared with surgery alone insquamous-cell cancer of the esophagus. N Engl J Med1997;337:1617.Burmeister BH, Smithers BM, Gebski V, et al. Surgery aloneversus chemoradiotherapy followed by surgery for resect-able cancer of the oesophagus: a randomised controlledphase III trial. Lancet Oncol 2005;6:65968.


    SUHAS V. PRADHAN (Syracuse, NY): Some years agore was a paper from Valerie Rusch & Associates frommorial. The value of PET [positron emission tomography]n in the workup meaning that if patients who were thought torelatively early cancer by ultrasound and other criteria, theynd that the SUV [standardized uptake value] was more than. In hindsight, they thought they would have fared muchter with neoadjuvant therapy followed by surgery.nd for the most part, we have been doing that, but I haven occasionally there have been bulky tumors with low PET

    n numbers, so in those patients I think ultrasound is of value.herwise, routinely Im not sure ultrasound is an absolute

    juvasease-free survival in patients with locally advanced rectalcancer. Ann Surg Oncol 2008;15:26617.Delanian S, Lefaix JL. Current management for late normaltissue injury: radiation-induced fibrosis and necrosis. SeminRadiat Oncol 2007;17:99107.Dolinsky CM, Mahmoud NN, Mick R, et al. Effect of timeinterval between surgery and preoperative chemoradiother-apy with 5-fluorouracil or 5-fluorouracil and oxaliplatin onoutcomes in rectal cancer. J Surg Oncol 2007;96:20712.Kerr SF, Norton S, Glynne-Jones R. Delaying surgery afterneoadjuvant chemoradiotherapy for rectal cancer may re-duce postoperative morbidity without compromising prog-nosis. Br J Surg 2008;95:153440.Lim SB, Choi HS, Jeong SY, et al. Optimal surgery time afterpreoperative chemoradiotherapy for locally advanced rectalcancers. Ann Surg 2008;248:24351.Moore HG, Gittleman AE, Minsky BD, et al. Rate of patho-logic complete response with increased interval betweenpreoperative combined modality therapy and rectal cancerresection. Dis Colon Rectum 2004;47:27986.Phlman L. Optimal timing of surgery after preoperativechemoradiotherapy for rectal cancer. Nat Clin Pract Oncol2009;6:1289.Tran CL, Udani S, Holt A, Arnell T, Kumar R, Stamos MJ.Evaluation of safety of increased time interval betweenchemoradiation and resection for rectal cancer. Am J Surg2006;192:8737.Courrech Staal EF, Aleman BM, Boot H, et al. Systematicreview of the benefits and risks of neoadjuvant chemoradia-tion for oesophageal cancer. Br J Surg 2010;97:148296.Ruol A, Rizzetto C, Castoro C, et al. Interval betweenneoadjuvant chemoradiotherapy and surgery for squamouscell carcinoma of the thoracic esophagus. Ann Surg 2010;252:78896.Swisher SG, Wynn P, Putnam JB, et al. Salvage esophagec-tomy for recurrent tumors after definitive chemotherapy andradiotherapy. J Thorac Cardiovasc Surg 2002;123:17583.

    cessary part of the workup if you are using a PET scan. Whatyour thoughts on that?

    KIM: The practice at our institution is to stage all patientsh both PET-CT [computed tomography] and endoscopicrasound if we are planning on esophagectomy.

    PRADHAN: Do you find that it gives you some additionalormation, ultrasound with the PET scan? I mean, you areking a decision about giving neoadjuvant or not giving neoad-Another possible weakness in the study is the inherents in the selection of patients with delayed surgery. Wenot include data on adverse events from CXRT;

    for rectal cancer: the Lyon R90-01 randomized trial. J ClinOncol 1999;17:2396.

    9. Tulchinsky H, Shmueli E, Figer A, Klausner JM, Rabau M.oadjuvant therapy. We did not include patients whoimately did not go on to surgery because of metastaticease, poor performance status, or patient choice. It istainly possible that some patients with lower perfor-nce status were found to have had a complete clinicalponse on restaging studies and chose not to undergorgery. We also excluded salvage cases. Data from ourperience suggest that salvage operations have a higherrbidity [19]. Including salvage cases would also obvi-sly decrease the rate of pCR in the delayed group.Nygaard K, Hagen S, Hansen HS, et al. Pre-operative radio-therapy prolongs survival in operable esophageal carci-noma: a randomized, multicenter study of pre-operativeradiotherapy and chemotherapy. The second Scandinaviantrial in esophageal cancer. World J Surg 1992;16:110410.Urba SG, Orringer MB, Turrisi A, Iannettoni M, Forastiere A,Strawderman M. Randomized trial of preoperative chemo-radiation versus surgery alone in patients with locoregionalesophageal carcinoma. J Clin Oncol 2001;19:30513.Francois Y, Nemoz CJ, Baulieux J, et al. Influence of theinterval between preoperative radiation therapy and surgeryon downstaging and on the rate of sphincter-sparing surgeryant; am I correct? Or are you giving neoadjuvant to all comers,some have proposed, except for very early intramucosal cancer?

  • DRKIM: The practice at our institution has evolved over the lastfew years and has moved towards adopting neoadjuvant chemo-radiation for the majority of tumors that are stage II and beyond.And for very early stage tumors, more patients are undergoingendoscopic mucosal resection. The PET/CT gives informationabout distant metastatic disease, whereas the EUS [endoscopicultrasound] is more accurate about locoregional disease, such asthe T stage of tumor and the involvement of regional lymphnodes.

    DR MARK KRASNA (Towson, MD): Dr Kim, thank you. Thatwas an excellent presentation. Just two quick questions.I would agree with you that especially when you are talking

    about neoadjuvant chemotherapy and radiation therapy, that itis appropriate to wait longer than the typical eight weeks, so Ithink your contribution is really important.I would ask one question that relates to the fact that since you

    held off on doing surgery for some of these patients until theyrecovered, I was a little surprised to see that there were so manypeople who were still skinny after you waited more than eightweeks. Could you talk about how you perhaps increased nutri-tion intake on the patients when you made the decision to delay?The second question, from the radiation oncology literature,

    both from basic research and also from clinical data, we knowthat especially in esophageal cancer but also in lung cancer, youwill continue to get tumor kill after high-dose radiation as muchas 12 weeks out. Again, I was interested to see that there was nodifference in path CR [pathologic complete response]. Could youcomment why you dont think there was an increased path CRrate even though some of your patients had greater than 12weeks evolved. I enjoyed your paper.

    DR KIM: Thank you. In terms of the question about nutrition, Ithink there was a difference in the baseline BMI [body massindex] between the groups, which goes along with the highernumber of squamous patients in the delayed-group, but weagree that there could have been more patients with poornutrition in the delayed group. Weight loss was a predictor forpostoperative complications in our study and its our practice todelay surgery for patients who are malnourished. Typically, aspatients recovered from the radiation induced esophagitis theywere able to swallow well and improve their nutritional statuswithout the need for adjunctive procedures. Occasionally, it wasnecessary to use a temporary feeding tube in an outpatientsetting, but this is typically reserved for patients that wereunderweight and had complaints of dysphagia and-or anorexia.So, I believe that you have made an excellent point; that clearly,based on our results and those of others we should be aggressiveabout optimizing nutrition prior to taking patients to surgery.And as far as the question about the rates of pathologic

    complete response, well, one thing is that we did actually do aseparate subgroup analysis looking at the patients who hadsurgery 12 weeks after and beyond and compared that withpatients who had surgery within 8 weeks, and there was still nodifference. However, this is a retrospective analysis includingonly those patients who went to surgery. It is likely that somepatients with lower performance status were found to have acomplete clinical response on restaging and then were advised,or chose to avoid surgery. We looked at our data in many ways,and really it may not be possible to resolve the issue ofdifferences in pathologic complete response rates based on aretrospective study. However, when one looks at the groups,with or without a denominator, the patients that did not go tosurgery do not have pathologic confirmation of response.

    213Ann Thorac Surg KIM ET AL2012;93:20713 TIMING OF ESOPHAGECTOMY



    Does the Timing of Esophagectomy After Chemoradiation Affect Outcome?Patients and MethodsSubgroup AnalysisStatistical Analysis

    ResultsSubgroup Analyses



View more >