Accepte

ReprintAve, CSsurgery.

0039-60

� 2012

doi:10.1

382 S

Impact of neoadjuvant chemotherapyon wound complications after breastsurgeryMarquita R. Decker, MD, MPH, David Y. Greenblatt, MD, MPH, Jeff Havlena, MS,Lee G. Wilke, MD, Caprice C. Greenberg, MD, MPH, and Heather B. Neuman, MD, MS, Madison, WI

Background. Use of neoadjuvant chemotherapy for breast cancer is increasing. The objective was toexamine risk of postoperative wound complications in patients receiving neoadjuvant chemotherapy forbreast cancer.Methods. Patients undergoing breast surgery from 2005 to 2010 were selected from the American Collegeof Surgeons National Surgical Quality Improvement Program database. Patients were included ifpreoperative diagnosis suggested malignancy and an axillary procedure was performed. We performed astepwise multivariable regression analysis of predictors of postoperative wound complications, overalland stratified by type of breast surgery. Our primary variable of interest was receipt of neoadjuvantchemotherapy.Results. Of 44,533 patients, 4.5% received neoadjuvant chemotherapy. Wound complications wereinfrequent with or without neoadjuvant chemotherapy (3.4% vs 3.1%; P = .4). Smoking, functionaldependence, obesity, diabetes, hypertension, and mastectomy were associated with wound complications.No association with neoadjuvant chemotherapy was seen (odds ratio [OR], 1.01; 95% confidence in-terval [CI], 0.78–1.32); however, a trend was observed toward increased complications in neoadjuvantpatients undergoing mastectomy with immediate reconstruction (OR, 1.58; 95% CI, 0.98–2.58).Conclusion. Postoperative wound complications after breast surgery are infrequent and not associatedwith neoadjuvant chemotherapy. Given the trend toward increased complications in patients undergoingmastectomy with immediate reconstruction, however, neoadjuvant chemotherapy should be among themany factors considered when making multidisciplinary treatment decisions. (Surgery 2012;152:382-8.)

From the Wisconsin Surgical Outcomes Research Program, Department of Surgery, University of WisconsinSchool of Medicine and Public Health, Madison, WI

INDICATIONS FOR NEOADJUVANT CHEMOTHERAPY IN THE

TREATMENT OF BREAST CANCER are expanding.1 Neoad-juvant chemotherapy provides the advantages ofmonitoring the in situ breast cancer for treatmentresponse and the potential for breast-conservingsurgery (BCS) in patients with locally advancedbreast cancer who otherwise may not have beencandidates.2 Given that both overall survival anddisease-free survival are equivalent after neoadju-vant and adjuvant chemotherapy, neoadjuvant che-motherapy can be considered as a treatmentoption for any patient who is expected to requiresystemic treatment.3-5 Therefore, there is a needto improve understanding of the potential for

d for publication May 10, 2012.

requests: Heather B. Neuman, MD, MS, 600 HighlandC H4/726, Madison, WI 53792-7375. E-mail: neuman@wisc.edu.

60/$ - see front matter

Mosby, Inc. All rights reserved.

016/j.surg.2012.05.001

URGERY

postoperative complications in recipients of neo-adjuvant chemotherapy.

A recent study assessing postoperative morbidityafter breast cancer surgery demonstrated that themost frequent complication was wound infection.6

Given that neutropenia can occur with breastcancer chemotherapeutics,7 this observation hasraised concern that patients treated with neoadju-vant chemotherapy may be at increased risk forpostoperative complications. This topic was beenexamined in several single-institution series thatfocused primarily on mastectomy with or withoutimmediate reconstruction; in these studies, no in-crease in postoperative complications in patientstreated with neoadjuvant chemotherapy was identi-fied8-11; however, the conclusions of these retro-spective studies conducted primarily at academiccenters are limited by small sample size, exclusionof patients undergoing breast conservation, andlack of generalizability to community settings.

The objective of this study was to examine therisk of postoperative wound complications in

SurgeryVolume 152, Number 3

Decker et al 383

patients receiving neoadjuvant chemotherapy forbreast cancer using the American College of Sur-geons---National Surgical Quality ImprovementProgram (ACS-NSQIP) database.12 By using thisprospectively collected, multi-institutional dataset,we were able to examine the relationship betweenreceipt of neoadjuvant chemotherapy and postop-erative wound complications, stratified by type ofoperative procedure, across multiple institutions,and reflecting a variety of settings. These resultshave the potential to influence the choice and se-quencing of the multidisciplinary components ofbreast cancer treatment.

METHODS

Data. The ACS-NSQIP is a quality improvementprogram that provides risk-adjusted surgical out-comes data to participating hospitals. In 2010, theACS-NSQIP database included data contributed by258 academic and community hospitals through-out the United States. Operative cases are selectedusing a systematic sampling process to minimizebias in case selection; in any 8-day sampling cycle,case selection is limited to no more than 3 breastlumpectomies. Preoperative and intraoperativevariables are collected via chart abstraction andother methods by a trained Surgical Clinical Re-viewer. Thirty-day postoperative outcomes includecomplications, mortality, reoperation, and dura-tion of stay. A full description of the ACS-NSQIPprogram is available online.12

Patient cohort. We selected a cohort of patientswho underwent (1) BCS (CPT codes 19162, 19301,19160, 19300, 19140, 19120, 19125, 19126), (2)mastectomy (CPT codes 19307, 19306, 19305,19302, 19240, 19220, 19200, 19304, 19303, 19182,19180), or (3) mastectomy and reconstruction(CPT code 19340 or codes for mastectomy plus19340, 19342, 15734, 19350, 19357, 19360, 19361,19364,19366, 19367, 19368, or 19369) from the2005–2010 ACS-NSQIP Participant Use Files. Pa-tients were included if their preoperative diagno-ses included malignant neoplasm of the breast(ICD-9 diagnosis codes 174.0–174.9), neoplasm ofthe breast (239.3), personal history of breast can-cer (v.10.3), secondary breast cancer (198.81),ductal carcinoma in situ (233.0), breast lump ormass (611.72), or inflammatory breast disease(611.0).

To further limit our cohort to patients likely tohave malignant disease (thereby eligible for neo-adjuvant chemotherapy), patients were also re-quired to have undergone either a sentinellymph node biopsy (CPT codes 38500, 38505, or38525), or an axillary lymph node dissection (CPT

codes 38740, 38745 or codes for modified radicalor radical mastectomy); patients who did notundergo an axillary staging procedure at thetime of their breast operation were excluded.Finally, we excluded male breast cancer patientsand any patients with metastatic cancer at the timeof their breast cancer operation.

Outcomes. The primary outcome was woundcomplication. Wound complications were catego-rized as superficial surgical site infection, deepinfection (including deep surgical site infectionand organ space infection), and wound dehis-cence. Although reoperation is considered acomplication by the ACS-NSQIP, we excluded itfrom our study because it is impossible to differ-entiate reoperation for positive margins or apositive sentinel lymph node from a true compli-cation that requires return to the operatingroom.

Variables. Our primary explanatory variablewas receipt of neoadjuvant chemotherapy within30 days of the operation. We also evaluated anumber of preoperative factors including demo-graphics (age, race), comorbidities (diabetesmellitus, hypertension, congestive heart failure,chronic obstructive pulmonary disease, steroiduse, bleeding disorder, history of cardiac surgery,history of stroke or transient ischemic event, openor infected wound), laboratory values (whiteblood cell count, hematocrit, platelets), andother factors (smoking, alcohol use, functionalstatus, body mass index). The breast operationwas categorized as BCS, mastectomy, or mastec-tomy with immediate reconstruction. The axillarypart of the operation was categorized as eitheraxillary lymph node dissection or sentinel lymphnode biopsy; patients undergoing both proce-dures were included in the axillary lymph nodedissection group.

Data analysis. General summary statistics weregenerated to describe our cohort. A univariate anal-ysis compared demographics and preoperative andoperative variables between patients who did and didnot receive neoadjuvant chemotherapy. TheMantel-Haenszel chi-squared test and the Fisher’s exact testwere used to examine the relationship between thevariables and the 2 chemotherapy groups. Variableshaving P # .1 were considered for inclusion into amultivariable logistic regressionmodel to further ex-plore the association between chemotherapy andwound complications while adjusting for other fac-tors. Multivariable models were built for the overallcohort, as well as stratified by type of breast opera-tion. All statistical analyses were performed usingSAS v.9.2 (Cary, NC).

SurgerySeptember 2012

384 Decker et al

RESULTS

A cohort of 44,533 patients was identified. Ofthese, 2,006 (4.5%) received neoadjuvant chemo-therapy. Patient demographics and preoperativeand operative characteristics are presented inTable I. Patients treated with neoadjuvant chemo-therapy were younger and had fewer comorbid-ities; however, they were more likely to smokeand to be overweight or obese. Additionally, neo-adjuvant patients had more frequent use of ste-roids and hematologic laboratory abnormalities.

Patients undergoing BCS were less likely to havebeen treated with neoadjuvant chemotherapy thanthose undergoing mastectomy (1.6% vs 5.8%; P <.001). Additionally, among those patients undergo-ing mastectomy, patients treated with neoadjuvantchemotherapy were less likely to receive immedi-ate reconstruction (21.3% vs 29.3%; P < .001). Ax-illary lymph node dissections were more commonin the neoadjuvant patient cohort (85.7% vs48.2%; P < .001).

As expected, short-term survival for patients wasexcellent, with an overall 30-day mortality of 0.06%(Table II). The 30-day morbidity was slightly less inpatients receiving neoadjuvant chemotherapy (5.6vs 7.0%; P = .01), but the wound complication ratewas comparable between the 2 groups (3.4% vs3.1%; P = .4), with a slightly greater rate of dehis-cence in the neoadjuvant chemotherapy cohort(0.7% vs 0.3%; P = .009).

In the multivariable analysis, factors predictiveof wound complications included smoking, func-tional dependence, overweight or obese, diabetes,hypertension, and mastectomy (Table III). Neoad-juvant chemotherapy was not associated withwound complication on multivariable analysis(odds ratio, 1.01; 95% confidence interval, 0.78–1.32; P = .9).

Because we anticipated that factors associatedwith wound complications would differ based onthe operative procedure performed, we performeda stratified analysis (Table IV). Patients undergoingBCS had a wound complication rate of 1.9%. Inthese patients, being overweight or obese, beingfunctionally dependent, and undergoing an axil-lary lymph node dissection was associated withwound complications. Compared with patients un-dergoing BCS, patients undergoing mastectomyhad a greater rate of wound complications(3.50% without reconstruction and 3.95% withimmediate reconstruction). Factors associatedwith an increased risk of wound complicationsincluded smoking, being overweight or obese,and diabetes. For patients not undergoing

immediate reconstruction, functional dependenceand hypertension were also associated with in-creased risk of wound complication. Receipt ofneoadjuvant chemotherapy was not significantly as-sociated with an increased risk of wound complica-tions, regardless of the operation performed. Atrend toward increased wound complications, how-ever, was seen in patients undergoing mastectomywith immediate reconstruction (odds ratio, 1.58;95% confidence interval, 0.98–2.58; P = .06).

DISCUSSION

In this analysis of the ACS-NSQIP database, weconfirmed the findings from prior studies dem-onstrating that wound complications after breastcancer surgery are infrequent (#4%). Woundcomplication rates vary by type of operative pro-cedure performed, with the least rates observedafter BCS (1.9%) and the greatest after mastec-tomy with reconstruction (4.0%). As expected,factors associated with wound complications alsovaried by type of operative procedure. For pa-tients undergoing BCS, age, obesity, functionaldependence, and undergoing a concurrent axil-lary lymph node dissection were associated withincreased risk. For patients undergoing mastec-tomy, factors differed. Some of the observedfactors, such as obesity, smoking, and hyperten-sion, have been noted in prior studies.6,13 We alsoidentified functional dependence and diabetes asincreasing risk; these factors have not beenreported previously. Regardless of the type of op-eration performed, use of neoadjuvant chemo-therapy did not significantly increase the risk ofwound complications.

Concern regarding the theoretic increased riskof wound complications after the receipt of neo-adjuvant chemotherapy has existed within thebreast surgery community for some time. In re-sponse to early concerns, the MD Anderson Can-cer Center developed a protocol regarding timingof operation for recipients of preoperative chemo-therapy. Under their ‘‘strict operative criteria,’’ allpatients treated with neoadjuvant chemotherapyhad to have a resectable neoplasm, white bloodcell count of >2,500 cells/mm3, and platelet countof >50,000 cells/mm3 before mastectomy.9 Thisearly, single-institution study demonstrated thatpostmastectomy morbidity in patients who re-ceived preoperative chemotherapy was no differ-ent from patients who did not have preoperativechemotherapy. Since that time, several studieshave reported similar findings8-11,14; in these stud-ies, ‘‘strict operative criteria’’ as reported by the

Table I. Patient demographic preoperative and operative characteristics

Variable

Neoadjuvant chemotherapy(n = 2,006)

No neoadjuvant chemotherapy(n = 42,527)

Pn % n %

DemographicsAge in years [mean (SD)] 52.1 (12.0) 59.63 (13.1) <.0001Race/ethnicityWhite 1,362 67.9 30,663 72.1 <.0001Black 256 12.8 4,314 10.1Other 224 11.2 3,874 9.1Missing 164 8.2 3,676 8.6

Preoperative health and comorbiditiesCurrent smoker 323 16.1 6,114 14.4 .032Alcohol use (>2 drinks/d) 13 0.7 636 1.5 .002Functional status: partial/fully dependent 22 1.1 582 1.4 .304

Body mass index classificationNormal weight 596 29.7 1,3842 32.6 .001Underweight 41 2.0 998 2.4Overweight/obese 1,367 68.2 27,536 64.8

Diabetes mellitus 111 5.5 3,726 8.8 <.0001Chronic obstructive pulmonary disease 24 1.2 1,133 2.7 <.0001Previous cardiac surgery 13 0.7 768 1.8 .0001Congestive heart failure 8 0.4 77 0.2 .029Hypertension (requiring medication) 602 30.0 18,209 42.8 <.0001History of transient ischemic attack 15 0.8 773 1.8 .0004Steroid use 79 3.9 515 1.2 <.0001Bleeding disorder 55 2.7 675 1.6 <.0001Open or infected wound 31 1.6 272 0.6 <.0001Preoperative lab values

White blood cell count >11 or <4.5 650 32.4 4,387 10.3 <.0001Missing 141 7.0 7,513 17.7

Hematocrit <38 1473 73.4 11,128 26.2 <.0001Missing 115 5.7 6,707 15.8

Platelet count <150 or >400 226 11.3 2,258 5.3 <.0001Missing 146 7.3 7,565 17.8

Operative variablesAmerican Society of Anesthesiologists classNo disturbance 48 2.4 3,029 7.1 <.0001Mild disturbance 1,234 61.5 26,898 63.3Severe disturbance 706 35.2 12,057 28.4Life-threatening disturbance 16 0.8 499 1.2

Blood transfusion 15 0.8 112 0.3 .003Wound classClean 1,930 96.2 41,574 97.8 <.0001Clean-contaminated/dirty/infected 76 3.8 953 2.2

Type of breast operationBreast-conserving surgery 220 11.0 13,571 31.9 <.0001Mastectomy 1,406 70.1 20,482 48.2Mastectomy with immediate reconstruction 380 18.9 8,474 19.9Axillary lymph node dissection 1,720 85.7 20,477 48.2 <.0001Sentinel lymph node biopsy 286 14.3 22,050 51.9

SD, Standard deviation.

SurgeryVolume 152, Number 3

Decker et al 385

MD Anderson group were not required. The find-ings of our study corroborate the conclusions ofthese single-institution, retrospective studies.

Recently, there has been growing interest inoutcomes after neoadjuvant chemotherapy andmastectomy with immediate reconstruction. Two

Table II. Frequency of postoperative complications, with and without neoadjuvant chemotherapy

Outcome variable

Neoadjuvant chemotherapy(n = 2,006)

No neoadjuvant chemotherapy(n = 42,527)

Pn % n %

30-Day mortality 4 0.20 23 0.05 .03230-Day morbidity 112 5.6 2,989 7.0 .013Overall wound complication 60* 3.4 1,304* 3.1 .413

Superficial infection 37 1.8 851 2.0 .624Dehiscence 13 0.65 119 0.28 .009Deep infection/abscess 20 1.0 376 0.88 .544

*Some patients experienced >1 type of wound complication.

Table III. Multivariable model of factors associated with wound complications in patients undergoingbreast cancer surgery

Factor n (%; n = 44,533) Adjusted odds ratio 95% Confidence interval P

Neoadjuvant chemotherapy 2,006 (4.5) 1.00 0.77–1.31 .973Age (yrs) 1.00 0.99–1.00 .252Race

White 32,025 (71.9) 1.00 ReferenceBlack 4,570 (10.3) 0.82 0.65–0.98 .431Other 4,098 (9.2) 0.80 0.65–0.98 .229

Smoker 6,437 (14.5) 1.56 1.35–1.80 <.0001Partial/fully dependent 604 (1.4) 2.02 1.43–2.88 <.0001Overweight/obese 28,903 (64.9) 2.16 1.87–2.50 <.0001Diabetes mellitus 3,837 (8.6) 1.57 1.33–1.85 <.0001Hypertension 18,811 (42.2) 1.25 1.10–1.43 .001Type of breast surgery

Breast-conserving surgery 13,791 (31.0) 1.00 ReferenceMastectomy 21,888 (49.2) 1.82 1.54–2.16 .034Mastectomy plus reconstruction 8,854 (19.9) 2.51 2.09–3.02 <.0001

Axillary lymph node dissection* 22,197 (49.8) 1.00 0.88–1.14 .945

*Analysis performed using sentinel lymph node biopsy as the reference.

SurgerySeptember 2012

386 Decker et al

single-institution studies of patients who under-went mastectomy and immediate reconstructionfound no difference in complication rates betweenpatients who received neoadjuvant chemotherapyand those who did not.8,11 Although we did notidentify an association between receipt of neoadju-vant chemotherapy and wound complications aftermastectomy with immediate reconstruction (P =.06), a trend was observed. This observation maybe noteworthy given the selection bias apparentin our patients undergoing immediate reconstruc-tion after neoadjuvant chemotherapy. Patientsselected for immediate reconstruction after neoad-juvant chemotherapy were younger and overall‘‘healthier’’ than those who underwent mastectomyalone, which may be a direct result of the chemo-therapy received. Other clinical data unavailablein the ACS-NSQIP dataset, including the stage ofthe cancer and receipt of postmastectomy

radiation, may also impact the risk of wound com-plications. Given the available data, we cannot con-clude that expanding the use of immediatereconstruction to all patients who undergo neoad-juvant chemotherapy would be acceptable. Ourdata support the conclusion that recipients of neo-adjuvant chemotherapy selected by their surgeonsto be good candidates for immediate reconstruc-tion can be expected to have good postoperativeoutcomes.

Several limitations exist for our study. First,patients in our study were not randomized toeither timing of chemotherapy or to type of oper-ation performed; as a result, selection bias exists.Differences between the patient groups were evi-dent for a number of factors, including age andcomorbidities; multivariable logistic regression wasused to adjust for this. Other variables relevant topatient selection, such as the stage of the cancer,

Table IV. Stratified Multivariable model of factors associated with wound complications in patientsundergoing breast cancer surgery

Factor n (%) Adjusted odds ratio 95% Confidence interval P

Breast-conserving surgery (n = 13,791)Neoadjuvant chemotherapy 220 (1.6) 0.43 0.10–1.75 .236Age — 0.99 0.98–1.00 .032RaceWhite 10,232 (74.2) 1.00 ReferenceBlack 1,320 (9.6) 1.22 0.84–1.78 .363Other 1,047 (7.6) 1.02 0.64–1.63 .739

Smoker 1,849 (13.4) 1.12 0.78–1.61 .536Partial/fully dependent 98 (0.7) 3.16 1.25–7.95 .015Overweight/obese 9,299 (67.4) 2.14 1.51–3.01 .0001Diabetes mellitus 1,107 (8.0) 1.33 0.89–2.00 .167Hypertension 6,202 (45.0) 1.14 0.85–1.53 .398Axillary lymph node dissection* 1,663 (12.1) 1.56 1.10–2.20 .012

Mastectomy (n = 21,888)Neoadjuvant chemotherapy 1,406 (6.4) 0.94 0.68–1.30 .706Age — 1.00 0.99–1.00 .242RaceWhite 15,124 (69.1) 1.00 ReferenceBlack 2,620 (12.0) 0.74 0.59–0.94 .254Other 2,317 (10.6) 0.74 0.56–0.97 .285

Smoker 3,368 (15.4) 1.68 1.39–2.02 <.0001Partial/fully dependent 481 (2.2) 2.01 1.37–2.97 .0004Overweight/obese 14,737 (67.3) 1.86 1.53–2.26 <.0001Diabetes mellitus 2,401 (11.0) 1.60 1.30–1.96 <.0001Hypertension 10,483 (47.9) 1.32 1.11–1.58 .002Axillary lymph node dissection* 15,992 (73.1) 0.90 0.76–1.58 .245

Mastectomy with immediate reconstruction (n = 8,854)Neoadjuvant chemotherapy 380 (4.3) 1.58 0.98–2.58 .062Age — 1.01 1.00–1.02 .209RaceWhite 6,669 (75.3) 1.00 ReferenceBlack 630 (7.1) 0.76 0.51–1.15 .510Other 734 (8.3) 0.78 0.51–1.21 .652

Smoker 1,220 (13.8) 1.60 1.21–2.14 .001Partial/fully dependent 25 (0.3) 0.90 0.12–6.94 .921Overweight/obese 4,867 (55.0) 2.86 2.17–3.77 <.0001Diabetes mellitus 329 (3.7) 1.86 1.22–2.84 .004Hypertension 2,126 (24.0) 1.16 0.89–1.52 .263Axillary lymph node dissection* 4,542 (51.3) 1.01 0.80–1.26 .949

*Analysis performed using sentinel lymph node biopsy as the reference.

SurgeryVolume 152, Number 3

Decker et al 387

were not available in our dataset and remainunaccounted for in this analysis.

Additionally, the sampling strategy utilized bythe NSQIP undersamples patients undergoingBCS compared with mastectomy; this fact is re-flected in our BCS rate of 31%, which is less thanwhat has been reported for the United Statespopulation.15,16 Because we are not reporting onrates of surgery, but rather complications accord-ing to surgery type, this difference should notimpact the conclusions of our study. Similarly,our reported rates of wound complications areless than complication rates reported in other

series,8,10 especially after breast reconstruction;this fact relates likely to the definitions of compli-cations used. Because the ACS-NSQIP is a qualityimprovement program designed largely to evaluateoutcomes after general and vascular surgery,disease-specific complications that would be rele-vant to breast cancer surgery (such as seromaaspiration or delay of chemotherapy) are not col-lected, which likely explains the difference be-tween the rates reported in our study and othersingle-institution series.

Next, it is possible that our cohort of neo-adjuvant chemotherapy recipients is incomplete.

SurgerySeptember 2012

388 Decker et al

In the ACS-NSQIP, chemotherapy received in the30 days before operation is recorded and thisvariable was used to define our neoadjuvant co-hort. It is possible that some neoadjuvant patientshad a greater interval between the end of chemo-therapy and operation, and therefore may havebeen misclassified. Currently, no standard existsfor the optimal interval between chemotherapyand operation. One recently published, random-ized, controlled trial described breast surgery be-tween 14 and 28 days after chemotherapy,2 andanother study reported a median 27-day interval,9

suggesting by comparison that the majority ofour cohort was categorized appropriately.

Finally, we were underpowered to detect smalldifferences in wound complication rates; >60,000neoadjuvant chemotherapy patients would havebeen required to achieve a power of 80%. Woundcomplication rates were low, even for the groupwho underwent mastectomy with reconstruction.Therefore, the authors believe that a clinicallymeaningful difference is unlikely to be observedwith additional patients.

Although limitations exist to our study, theseresults represent currently the most comprehen-sive evaluation of neoadjuvant chemotherapy andpostoperative wound complications after the oper-ative treatment of local-regional breast cancer. Byusing the ACS-NSQIP database, we were able toexamine the outcomes from a multi-institutionalcohort representing a range of community andacademic surgical practices. Additionally, we hadthe benefit of prospectively collected and validatedpreoperative and operative variables, which al-lowed us to control for both previously describedrisk factors for wound complications (includingobesity, smoking, axillary dissection, and mastec-tomy),6,13,17,18 as well as other unique factors. Oneof the greatest strengths of our study, however, isthe rigorous methodology applied to collectionof postoperative complications by the ACS-NSQIP. This increases the reliability and validityof our conclusions.

REFERENCES

1. Kaufmann M, von Minckwitz G, Bear HD, Buzdar A, McGaleP, Bonnefoi H, et al. Recommendations from an interna-tional expert panel on the use of neoadjuvant (primary) sys-temic treatment of operable breast cancer: new perspectives2006. Ann Oncol 2007;18:1927-34.

2. Loibl S, von Minckwitz G, Raab G, Blohmer J-U, Dan CostaS, Gerber B, et al. Surgical procedures after neoadjuvantchemotherapy in operable breast cancer: results of the GE-PARDUO trial. Ann Surg Oncol 2006;13:1434-42.

3. Rastogi P, Anderson SJ, Bear HD, Geyer CE, KahlenbergMS, Robidoux A, et al. Preoperative chemotherapy: updatesof National Surgical Adjuvant Breast and Bowel Project Pro-tocols B-18 and B-27. J Clin Oncol 2008;26:778-85.

4. Waljee JF, Newman LA. Neoadjuvant systemic therapy andthe surgical management of breast cancer. Surg Clin NorthAm 2007;87:399-415.

5. Mauri D, Polyzos NP, Salanti G, Pavlidis N, Ioannidis JPA.Multiple-treatments meta-analysis of chemotherapy and tar-geted therapies in advanced breast cancer. J Natl CancerInst 2008;100:1780-91.

6. El-Tamer MB, Ward BM, Schifftner T, Neumayer L, Khuri S,Henderson W, et al. Morbidity and mortality followingbreast cancer surgery in women: national benchmarks forstandards of care. Ann Surg 2007;245:665-71.

7. Bear HD, Anderson S, Brown A, Smith R, Mamounas EP,Fisher B, et al. The effect on tumor response of adding se-quential preoperative docetaxel to preoperative doxorubi-cin and cyclophosphamide: preliminary results fromNational Surgical Adjuvant Breast and Bowel Project Proto-col B-27. J Clin Oncol 2003;21:4165-74.

8. Warren Peled A, Itakura K, Foster RD, Hamolsky D, TanakaJ, Ewing C, et al. Impact of chemotherapy on postoperativecomplications after mastectomy and immediate breast re-construction. Arch Surg 2010;145:880-5.

9. Broadwater JR, Edwards MJ, Kuglen C, Hortobagyi GN,Ames FC, Balch CM, et al. Mastectomy following preopera-tive chemotherapy. Strict operative criteria control opera-tive morbidity. Ann Surg 1991;213:126-9.

10. Hu Y-Y, Weeks CM, In H, Dodgion CM, Golshan M, ChunYS, et al. Impact of neoadjuvant chemotherapy on breast re-construction. Cancer 2011;117:2833-41.

11. Azzawi K, Ismail A, Earl H, Forouhi P, Malata CM. Influ-ence of neoadjuvant chemotherapy on outcomes of im-mediate breast reconstruction. Plast Reconstr Surg 2010;126:1-11.

12. AmericanCollege of Surgeons–National SurgicalQuality Im-provement Initiative. Available from: http://www.acsnsqip.org. Accessed December 31, 2011.

13. McCarthy CM, Mehrara BJ, Riedel E, Davidge K, Hinson A,Disa JJ, et al. Predicting complications following expander/implant breast reconstruction: an outcomes analysis basedon preoperative clinical risk. Plast Reconstr Surg 2008;121:1886-92.

14. Deutsch MF, Smith M, Wang B, Ainsle N, Schusterman MA.Immediate breast reconstruction with the TRAM flap afterneoadjuvant therapy. Ann Plast Surg 1999;42:240-4.

15. Greenberg CC, Lipsitz SR, Hughes ME, Edge SB, TheriaultR, Wilson JL, et al. Institutional variation in the surgicaltreatment of breast cancer: a study of the NCCN. AnnSurg 2011;254:339-45.

16. Morrow M, Jagsi R, Alderman AK, Griggs JJ, Hawley ST,Hamilton AS, et al. Surgeon recommendations and receiptof mastectomy for treatment of breast cancer. JAMA 2009;302:1551-6.

17. Khansa I, Colakoglu S, Curtis MS, Yueh JH, Ogunleye A, To-bias AM, et al. Postmastectomy breast reconstruction afterprevious lumpectomy and radiation therapy: analysis ofcomplications and satisfaction. Ann Plast Surg 2011;66:444-51.

18. Nahabedian MY, Tsangaris T, Momen B, Manson PN. Infec-tious complications following breast reconstruction withexpanders and implants. Plast Reconstr Surg 2003;112:467-76.