Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2011 Apr 19.
Published in final edited form as: Cancer. 2009 Sep 15;115(18):4104–4109. doi: 10.1002/cncr.24466

Quality of pathologic response and surgery correlate with survival for completely resected bladder cancer following neoadjuvant chemotherapy

Guru Sonpavde 1, Bryan H Goldman 2, VO Speights 3, Seth P Lerner 1, David P Wood 4, Nicholas J Vogelzang 5, Donald L Trump 6, Ronald B Natale 7, H Barton Grossman 8, E David Crawford 9
PMCID: PMC3079554  NIHMSID: NIHMS244858  PMID: 19517476

Abstract

BACKGROUND

In a retrospective study of SWOG-S8710/INT-0080 (radical cystectomy [RC] alone vs 3 cycles of MVAC neoadjuvant chemotherapy [NC] before RC for bladder cancer), factors associated with improved overall survival (OS) included pathologic complete response (pCR) defined as P0, treatment with NC, completion of RC with negative margins and ≥10 pelvic lymph nodes (LNs) removed.

METHODS

We used stratified Cox regression to retrospectively study the association of quality of pathologic response post-RC with OS in the subset of S8710 patients that received NC and RC with negative margins.

RESULTS

Of 154 patients who received NC, 68 (44.2%) were <P2 (P0, Pa, P1, or CIS) at RC, 46 (29.9%) were P0, and the remainder had P2+ disease or did not undergo RC. In 115 patients who had RC with negative margins, compared to P0 patients, those with residual Pa, P1 or CIS appeared to have worse OS (p=.054); OS was significantly worse for residual P2+ (p=.0006). LN+ disease was associated with worse OS than LN- (p=.0005). In patients with LN- disease, those with <10 nodes removed appeared to have inferior OS compared to 10+ nodes (p=.079). The combination of pre-NC clinical stage and post-RC pathologic stage was predictive of OS (p<.0001).

CONCLUSION

NC and RC with negative margins for bladder cancer followed by pathological P0 and LN- disease correlate with improved OS. A combination of baseline clinical stage and post-RC pathologic stage may better predict OS.

Keywords: Neoadjuvant chemotherapy, bladder cancer, pathologic complete response, overall survival

INTRODUCTION

Neoadjuvant MVAC (methotrexate, vinblastine, doxorubicin, and cisplatin) chemotherapy was evaluated for locally advanced bladder cancer (clinical T2 to T4a) treated with radical cystectomy (RC) in a phase III trial (S8710) conducted by SWOG (Southwest Oncology Group) [1]. A total of 317 patients were randomized to RC or 3 cycles of MVAC followed by RC. The median survival among patients assigned to surgery alone was 46 months compared to 77 months with MVAC (P=0.06). Significantly more patients in the MVAC group had no residual disease (P0) than patients in the cystectomy group (38% vs. 15%, P<0.001). In both groups, improved survival was associated with P0 disease, with 5-year survival for all patients with P0 being an impressive 85%. The International Collaboration of Trialists trial and a meta-analysis have confirmed the value of neoadjuvant cisplatin-based combination chemotherapy [2-4]. Retrospective analyses of S8710 revealed that survival was associated with neoadjuvant MVAC, completion of RC and ≥10 pelvic lymph nodes (LNs) removed [5,6]. Five year survival was 81% in patients who had neoadjuvant MVAC and RC with ≥10 LNs removed (n=66), 66% in patients given RC alone with ≥10 LNs removed (n=60), 55% for those given MVAC and RC with <10 LNs removed (N=49), and 39% for those given RC alone with <10 LNs removed (n=44). Pooled across treatment arms, the 5-year survival of patients with positive surgical margins (N=25) and those who did not undergo RC (N=39) was dismal at 0% and 11%, respectively. Therefore, both the quality of surgery and neoadjuvant chemotherapy appeared to impact favorably and independently on outcomes.

pCR (pathologic complete remission) with neoadjuvant therapy for breast cancer is also highly correlated with survival. However, the optimal definition of pCR is unsettled [7-12]. When there is no residual invasive cancer in the breast, the number of involved axillary lymph nodes is inversely related to survival. Conversely, patients who convert to node-negative status after treatment have excellent survival, even with residual disease in the breast. There is no evidence that residual in situ carcinoma increases risk of distant relapse [13,14]. In a recent retrospective analysis, residual cancer burden calculated following neoadjuvant chemotherapy for breast cancer as a continuous index combining primary tumor (size and cellularity) and nodal (number and size) characteristics for prediction of distant relapse-free survival was independently prognostic in a multivariate model [15]. In another retrospective analysis of women who received neoadjuvant chemotherapy, risk-stratification of patients was facilitated by the combination of pre-treatment clinical stage and post-therapy pathologic stage [16]. S8710 defined pCR as the absence of all residual bladder cancer (P0). However, residual non-muscle invasive cancer including CIS (carcinoma in situ), Pa and P1 disease may correlate with outcomes similar to P0 and may differentially impact prognosis compared to P2+ disease. Similar to breast cancer, outcomes following neoadjuvant chemotherapy are likely associated with pre-therapy clinical staging and post-therapy pathologic staging and tumor burden. Therefore, we planned a retrospective study of the S8710 trial to evaluate the correlation of the quality of pathologic response to neoadjuvant MVAC and surgical resection with OS.

MATERIALS AND METHODS

Retrospective cohort study of a prospective trial

We retrospectively studied patients in S8710 who received neoadjuvant MVAC chemotherapy and completed RC with negative margins. The primary goals were to determine the association of pathologic response (P0 vs Pa, P1, CIS vs P2+), pathologic lymph node involvement and quality of pelvic lymph node dissection with overall survival (OS). In addition, the correlation of a combination of baseline clinical stage and RC pathological stage with OS was analyzed.

Statistical analysis

Associations between various pathologic and clinical features in the subset of S8710 patients that received NC and RC with negative margins with overall survival were evaluated using stratified proportional hazards regression. Baseline clinical T stage (T2 vs T3-4), pathologic stage at cystectomy (P0 vs Pa/P1/CIS vs P2+), and number of lymph nodes removed (<10 vs 10+) were explored as potential prognostic factors. Since all patients included in the analysis received RC, Day 0 in these survival models was defined as the date of cystectomy.

RESULTS

Quality of pathologic response with neoadjuvant MVAC appears to be a function of baseline stage

Of all 154 patients who received neoadjuvant MVAC, 68 (44%) had <P2 disease (P0, Pa, P1, or CIS) at RC, and 46 (30%) had P0 disease. The remainder of patients exhibited residual ≥P2 disease or did not undergo RC for a variety of reasons. Among the 62 (40%) with clinical T2 disease at baseline, 34 (55%) and 24 (39%) were <P2 and P0, respectively. Among the 92 (60%) with T3-4 disease, 32 (35%) were <P2 and 22 (24%) were P0.

Pathologic stage at cystectomy confers prognostic impact

In the 115 of these 154 patients who underwent RC with negative margins, several pathologic features were associated with OS (Table 1). Pathologic response defined as either P0 or <P2 predicted for enhanced OS. The hazard ratio (HR) of OS for >P0 versus P0 was 2.51 (95% CI=1.47, 4.27; p=0.0008), while the HR for P2+ versus <P2 was 2.24 (95% CI=1.34, 3.76; p=0.0022). Compared to P0 patients, those with residual Pa, P1 or CIS appeared to have worse OS (Figure 1), but the difference did not quite achieve statistical significance (HR (95% CI)=2.05 (0.99, 4.24); p=.054), while OS was significantly worse for residual P2+ (HR (95% CI)=2.75 (1.54,4.89); p=.0006). Pathological lymph node (LN) positive disease was associated with worse OS than LN negative disease (HR (95% CI)=2.90 (1.59,5.28); p=.0005) (Figure 2) .

Table 1.

Correlation of clinicopathologic features with overall survival among patients treated with neoadjuvant MVAC and RC

Characteristic N Median OS (yrs) HR (95% CI)
Pathologic Stage
P0 46 13.6 1.0
P1/CIS/Pa 22 10.6 2.05 (0.99, 4.24)
P2+ 47 3.7 2.75 (1.54, 4.89)*
Node Status+
LN- 93 10.6 1.0
LN+ 17 2.4 2.90 (1.59, 5.28)*
Nodes Removed (LN- Only)
10+ 50 13.7 1.0
<10 43 7.2 1.66 (0.94, 2.90)
Clinical T Stage/Pathology
cT2 → <P2 35 NR 1.0
cT3-4 → <P2 33 8.9 1.87 (0.96, 3.63)
cT2 → P2+ 13 1.8 7.64 (3.51, 16.61)*
cT3-4 → P2+ 34 5.1 2.82 (1.48, 5.38)*
Open in a new tab

Index: OS overall survival, HR (hazard ratio) relative to the reference of 1.0

+

Nodal status missing for 5 patients

*

p<.005

Figure 1.

Figure 1

Open in a new tab

Postcystectomy survival by post-MVAC pathologic stage

Figure 2.

Figure 2

Open in a new tab

Survival based on pathological lymph node disease at cystectomy

The quality of lymph node dissection has prognostic value

In patients with LN negative disease, those with <10 nodes removed appeared to have inferior OS compared to 10+ nodes, although this result was not statistically significant (HR (95% CI)=1.66 (0.94,2.90); p=.079) (Figure 3) . The median OS was 7.2 years for <10 LNs removed versus 13.7 years for ≥ 10 LNs removed.

Figure 3.

Figure 3

Open in a new tab

Survival based on number of lymph nodes removed for pathological N0 disease

The combination of baseline clinical stage and post-MVAC pathologic stage as a prognostic factor

The combination of baseline clinical stage and post-RC pathologic stage (<P2 versus P2+) was prognostic for OS (Figure 4). Among patients with initial T2 disease, P2+ disease post-RC was associated with a >7-fold higher risk of death than <P2 (Table 1).

Figure 4.

Figure 4

Open in a new tab

Postcystectomy survival by baseline clinical stage and post-MVAC pathologic stage (<P2 versus P2+)

DISCUSSION

Among patients who received NC and RC with negative margins for bladder cancer, pathological P0 and LN- disease at RC correlate with improved OS. Compared to P0 patients, those with residual Pa, P1 or CIS appeared to have worse OS, while those with P2+ disease had significantly worse OS. A combination of baseline clinical stage and post-RC pathologic stage may better predict OS. Since positive margins and lack of RC correlated with dismal outcomes in a previous retrospective analysis of all patients enrolled on the S8710 trial, we conducted this analysis only in patients who underwent a complete RC with negative margins [5,6]. Although this retrospective analysis is limited due to the modest number of patients studied, it suggests that residual pathologic stage following neoadjuvant MVAC is associated with post-cystectomy survival in patients that undergo RC with negative margins. Progression-free survival (PFS) could not be analyzed due to the inadequate capture of this endpoint in the prospective trial. Furthermore, due to the very long follow-up time now available on these patients, a large number of non-cancer related deaths may have been captured in this analysis. Therefore, the hazard ratio estimates presented here may represent conservative estimates of the true association between survival and pathologic and clinical stage. Additionally, the original pathology slides and specimens were not available for a more detailed analysis of residual tumor volume. Central pathology review was performed for the initial tumor biopsy, but not for the cystectomy specimen. Among LN negative patients, those with more extensive LN dissection had superior outcomes, although this result was not statistically significant. Similar findings have been reported elsewhere in RC patients who did not receive neoadjuvant chemotherapy [17]. The prognostic impact of the number of lymph nodes removed in node positive patients was not analyzed since there were few patients in this category (n=17) and the dismal OS of patients with node positive disease was already known.

Patients with lower baseline clinical stage (T2) achieved subsequent major pathologic response (<P2) more often (73%) than those with T3-4 disease (49%). Baseline T2 disease followed by <P2 at RC was associated with the best post-RC survival, whereas those with T2 and subsequent residual muscle-invasive disease (P2+) had the worst outcomes, and may represent patients with chemoresistant disease, although this result is based upon a small subset (N=13) and therefore must be interpreted with caution. Tumor biology is probably the underlying factor driving the chemoresponsiveness in addition to initial clinical stage, but tumor genomics and proteomics could not be evaluated on this study. Since a proportion of patients with T2 disease may attain <P2 residual stage with the initial cystoscopic biopsy or transurethral resection of the bladder tumor (TURBT), an independent impact of maximal TURBT on OS may also be occurring in patients receiving MVAC; i.e. all of the benefit may not be attributable to neoadjuvant chemotherapy. Therefore, urologists should probably still perform an optimal and maximal TURBT in patients with T2 disease that may receive neoadjuvant chemotherapy. Studies support some breast cancers being generally sensitive to different classes of chemotherapy whereas others may be intrinsically resistant to different classes of chemotherapy [18,19]. This finding in the setting of breast cancer may be applicable to the bladder cancer setting. Our data are hypothesis-generating and with further validation, they may assist in optimally risk-stratifying patients enrolled in future trials of adjuvant therapy with potentially non-cross resistant biologic agents based on pathologic stage of residual disease following neoadjuvant chemotherapy.

While adjuvant chemotherapy has been evaluated for muscle-invasive bladder cancer, definitive data from large randomized trials have not been available [20-23]. Additionally, a recent retrospective study demonstrated that 30% of patients that underwent radical cystectomy suffered from complications within 90 days that could contraindicate adjuvant chemotherapy [24]. Response to neoadjuvant MVAC may also be predicted by genomic and proteomic analysis of tumor tissue obtained at baseline biopsy. Fourteen genes predicted pCR in a retrospective study of patients with bladder cancer who received neoadjuvant MVAC [25,26]. Among those genes, topoisomerase IIa (target of doxorubicin) was down-regulated in the nonresponder group. Correlation was found between low/intermediate BRCA1 mRNA levels (which mediates DNA repair) and pCR as well as long-term outcomes with neoadjuvant cisplatin-combination chemotherapy for bladder cancer in a retrospective study of 49 patients [27]. In the setting of neoadjuvant therapy of breast cancer, early modulation of pharmacodynamic biomarkers (e.g. Ki-67) after brief neoadjuvant therapy appears to correlate with long-term clinical outcomes and this paradigm needs to be evaluated in bladder cancer [28,29]. Thus, the neoadjuvant paradigm holds considerable promise to develop tailored therapy for patients with locally advanced bladder cancer. Additionally, this paradigm has the potential to enable the efficient use of resources and accelerate the pace of systemic therapy development due to the availability of pathologic response as an intermediate endpoint.

Acknowledgments

This investigation was supported in part by the following PHS Cooperative Agreement grant numbers awarded by the National Cancer Institute, DHHS: CA32102, CA38926, CA21115, CA35421, CA46441, CA22433, CA42777, CA58861, CA59416, CA46282, CA27057, CA14028, CA46113, CA20319, CA46136, CA45377, CA128567, CA45560, CA35431, CA32734, CA35261, CA35090, CA16385, CA58882, CA76447, CA46368, CA68183, CA28862, CA58415, CA35281, CA63844, CA35192, CA35117, CA35084, CA58686, CA35262

Footnotes

Disclosures:

Guru Sonpavde: Speakers’ bureau for Sanofi-Aventis, Wyeth, Pfizer and Novartis; Research support from Pfizer, Eli Lilly, BMS, Astrazeneca, Novartis, Celgene and Cytogen.

REFERENCES

  • 1.Grossman HB, Natale RB, Tangen CM, et al. Neoadjuvant chemotherapy plus cystectomy compared with cystectomy alone for locally advanced bladder cancer. N Engl J Med. 2003;349(9):859–66. doi: 10.1056/NEJMoa022148. Erratum in: N Engl J Med. 2003; 349 (19): 1880. [DOI] [PubMed] [Google Scholar]
  • 2.Neoadjuvant cisplatin, methotrexate, and vinblastine chemotherapy for muscle-invasive bladder cancer: a randomised controlled trial. International collaboration of trialists. Lancet. 1999;354(9178):533–40. Erratum in: Lancet 1999; 354 (9190): 1650. [PubMed] [Google Scholar]
  • 3.Hall RR, behalf of the Intl Collaboration of Trialists of the MRC Advanced Bladder Cancer Group, MRC Clinical Trials Unit, London, UK Updated results of a randomised controlled trial of neoadjuvant cisplatin (C), methotrexate (M) and vinblastine (V) chemotherapy for muscle-invasive bladder cancer. Proc Am Soc Clin Oncol. 2002 abstr No. 710. [Google Scholar]
  • 4.Advanced Bladder Cancer Meta-analysis Collaboration Neoadjuvant chemotherapy in invasive bladder cancer: a systematic review and meta-analysis. Lancet. 2003;361(9373):1927–34. doi: 10.1016/s0140-6736(03)13580-5. [DOI] [PubMed] [Google Scholar]
  • 5.Herr HW, Faulkner JR, Grossman HB, et al. Surgical factors influence bladder cancer outcomes: a cooperative group report. J Clin Oncol. 2004;22(14):2781–9. doi: 10.1200/JCO.2004.11.024. [DOI] [PubMed] [Google Scholar]
  • 6.Dotan ZA, Bajorin DF, Grossman BH, Herr HW. Optimal Combined Modality Treatment Improves Outcome of Locally Advanced Bladder Cancer: Analysis of SWOG 8710. J Clin Oncol 2005 ASCO Annual Meeting Proceedings. 2005 June 1;23(16S)(Supplement):4531. Part I of II. [Google Scholar]
  • 7.Kuerer HM, Newman LA, Smith TL, et al. Clinical course of breast cancer patients with complete pathologic primary tumor and axillary lymph node response to doxorubicin-based neoadjuvant chemotherapy. J Clin Oncol. 1999;17:460–469. doi: 10.1200/JCO.1999.17.2.460. [DOI] [PubMed] [Google Scholar]
  • 8.Bear HD, Anderson S, Smith RE, et al. Sequential preoperative or postoperative docetaxel added to preoperative doxorubicin plus cyclophosphamide for operable breast cancer: National Surgical Adjuvant Breast and Bowel Project Protocol B-27. J Clin Oncol. 2006;24:2019–2027. doi: 10.1200/JCO.2005.04.1665. [DOI] [PubMed] [Google Scholar]
  • 9.Kaufmann M, Hortobagyi GN, Goldhirsch A, et al. Recommendations from an international expert panel on the use of neoadjuvant (primary) systemic treatment of operable breast cancer: An update. J Clin Oncol. 2006;24:1940–1949. doi: 10.1200/JCO.2005.02.6187. [DOI] [PubMed] [Google Scholar]
  • 10.Kurosumi M. Significance of histopathological evaluation in primary therapy for breast cancer– recent trends in primary modality with pathological complete response (pCR) as endpoint. Breast Cancer. 2004;11:139–147. doi: 10.1007/BF02968293. [DOI] [PubMed] [Google Scholar]
  • 11.Kuroi K, Toi M, Tsuda H, et al. Unargued issues on the pathological assessment of response in primary systemic therapy for breast cancer. Biomed Pharmacother. 2005;59:S387–S392. doi: 10.1016/s0753-3322(05)80085-x. [DOI] [PubMed] [Google Scholar]
  • 12.von Minckwitz G, Raab G, Caputo A, et al. Doxorubicin with cyclophosphamide followed by docetaxel every 21 days compared with doxorubicin and docetaxel every 14 days as preoperative treatment in operable breast cancer: The GEPARDUO study of the German Breast Group. J Clin Oncol. 2005;23:2676–2685. doi: 10.1200/JCO.2005.05.078. [DOI] [PubMed] [Google Scholar]
  • 13.Jones RL, Lakhani SR, Ring AE, et al. Pathological complete response and residual DCIS following neoadjuvant chemotherapy for breast carcinoma. Br J Cancer. 2006;94:358–362. doi: 10.1038/sj.bjc.6602950. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Mazouni C, Peintinger F, Wan-Kau S, et al. Residual ductal carcinoma in situ in patients with complete eradication of invasive breast cancer after neoadjuvant chemotherapy does not adversely affect patient outcome. J Clin Oncol. 2007;25:2650–2655. doi: 10.1200/JCO.2006.08.2271. [DOI] [PubMed] [Google Scholar]
  • 15.Symmans WF, Peintinger F, Hatzis C, et al. Measurement of residual breast cancer burden to predict survival after neoadjuvant chemotherapy. J Clin Oncol. 2007;25(28):4414–22. doi: 10.1200/JCO.2007.10.6823. [DOI] [PubMed] [Google Scholar]
  • 16.Jeruss JS, Mittendorf EA, Tucker SL, et al. Combined use of clinical and pathologic staging variables to define outcomes for breast cancer patients treated with neoadjuvant therapy. J Clin Oncol. 2008;26(2):246–52. doi: 10.1200/JCO.2007.11.5352. [DOI] [PubMed] [Google Scholar]
  • 17.Poulsen A, Horn T, Steven K. Radical cystectomy: extending the limits of pelvic lymph node dissection improves survival for patients with bladder cancer confined to the bladder wall. J Urol. 1998;160:2015–19. [PubMed] [Google Scholar]
  • 18.Smith IE, Heys SD, Hutckeon AW, et al. Neoadjuvant chemotherapy in breast cancer: significantly enhanced response with docetaxel. J Clin Oncol. 2002;20(6):1456–1466. doi: 10.1200/JCO.2002.20.6.1456. [DOI] [PubMed] [Google Scholar]
  • 19.von Minckwitz G, Kummel S, Vogel P, et al. Neoadjuvant vinorelbine-capecitabine versus docetaxel-doxorubicin-cyclophosphamide in early nonresponsive breast cancer: phase III randomized GeparTrio trial. J Natl Cancer Inst. 2008;100(8):542–551. doi: 10.1093/jnci/djn085. [DOI] [PubMed] [Google Scholar]
  • 20.Skinner DG, Daniels JR, Russell CA, et al. The role of adjuvant chemotherapy following cystectomy for invasive bladder cancer: a prospective comparative trial. J Urol. 1991;145(3):459–64. doi: 10.1016/s0022-5347(17)38368-4. [DOI] [PubMed] [Google Scholar]
  • 21.Lehmann J, Franzaring L, Thuroff J, et al. Complete long-term survival data from a trial of adjuvant chemotherapy vs control after radical cystectomy for locally advanced bladder cancer. BJU Int. 2006;97(1):42–7. doi: 10.1111/j.1464-410X.2006.05859.x. [DOI] [PubMed] [Google Scholar]
  • 22.Freiha F, Reese J, Torti FM. A randomized trial of radical cystectomy versus radical cystectomy plus cisplatin, vinblastine and methotrexate chemotherapy for muscle invasive bladder cancer. J Urol. 1996;155(2):495–9. [PubMed] [Google Scholar]
  • 23.Studer UE, Bacchi M, Biedermann C, et al. Adjuvant cisplatin chemotherapy following cystectomy for bladder cancer: results of a prospective randomized trial. J Urol. 1994;152(1):81–4. doi: 10.1016/s0022-5347(17)32822-7. [DOI] [PubMed] [Google Scholar]
  • 24.Donat MS, Shabsigh A, Savage C, et al. The impact of postoperative complications on the timing of adjuvant chemotherapy in patients undergoing radical cystectomy for invasive or node positive bladder cancer.. Proc Amer Urol Assoc Annual Meeting 2008; Orlando, FL. abstract 848. [Google Scholar]
  • 25.Takata R, Katagiri T, Kanehira M, et al. Predicting response to methotrexate, vinblastine, doxorubicin, and cisplatin neoadjuvant chemotherapy for bladder cancers through genome-wide gene expression profiling. Clin Cancer Res. 2005;11(7):2625–36. doi: 10.1158/1078-0432.CCR-04-1988. [DOI] [PubMed] [Google Scholar]
  • 26.Takata R, Katagiri T, Kanehira M, et al. Validation study of the prediction system for clinical response of M-VAC neoadjuvant chemotherapy. Cancer Sci. 2007;98(1):113–7. doi: 10.1111/j.1349-7006.2006.00366.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Font A, Tarón M, Costa C, et al. BRCA1 mRNA expression in patients (p) with bladder cancer treated with neoadjuvant cisplatin-based chemotherapy. J Clin Oncol. 2008 May 20;26(suppl) abstr 5021. [Google Scholar]
  • 28.Dowsett M, Smith IE, Ebbs SR, et al. Prognostic value of Ki67 expression after short-term presurgical endocrine therapy for primary breast cancer. J Natl Cancer Inst. 2007;99(2):167–70. doi: 10.1093/jnci/djk020. [DOI] [PubMed] [Google Scholar]
  • 29.Jones RL, Salter J, Parton ANM, et al. Prognostic significance of Ki67 before and after neoadjuvant chemotherapy in early breast cancer. J Clin Oncol. 2007;25:21s. doi: 10.1007/s10549-008-0081-7. abstr 573. [DOI] [PubMed] [Google Scholar]

RESOURCES