Definition & Impact of PCR on Prognosis After Neoadjuvant Chemo in Various BC Types

[Hopeful]

J Clin Oncol. 2012 Apr 19;[Epub Ahead of Print], G von Minckwitz, M Untch, J-U Blohmer, SD Costa, H Eidtmann, PA Fasching, B Gerber, W Eiermann, J Hilfrich, J Huober, C Jackisch, M Kaufmann, GE Konecny, C Denkert, V Nekljudova, K Mehta, S Loibl

Abstract

This pooled analysis looked at pCR definitions and their prognostic impact in patients treated with neoadjuvant chemotherapy for various breast cancer subtypes. Results suggest that pCR is not an adequate surrogate endpoint for all subtypes.

TAKE-HOME MESSAGE

This pooled analysis looked at pCR definitions and their prognostic impact in patients treated with neoadjuvant chemotherapy for various breast cancer subtypes. Results suggest that pCR is not an adequate surrogate endpoint for all subtypes.

EXPERT COMMENTARY

Lee S. Schwartzberg, MD, FACP

Neoadjuvant chemotherapy in patients with breast cancer reduces the need for mastectomy and allows in vivo assessment of the benefit of chemotherapy. Pathologic complete response (pCR) is held as the gold standard for predicting long-term outcome. However, the field is hampered by multiple different definitions of pCR. Is pCR defined as no invasive tumor in the breast? No in situ tumor? No involvement of the lymph nodes? Or by other ways? And what does pCR mean in the context of our growing recognition that breast cancer is a group of different diseases, frequently defined as the intrinsic classification molecularly, which is approximated by knowing the ER, PR, HER2 status and the degree of proliferation?Von Minckwitz and colleagues have analyzed outcome by various pCR definitions in a large series of patients treated in sequential clinical trials investigating neoadjuvant therapy. They demonstrate that the definition of pCR does matter, in that the patients who are truly without evidence of disease in both the breast and lymph nodes have an excellent prognosis and can be distinguished from all other groups. Somewhat surprisingly, patients with residual in situ carcinoma only did not fare as well, and had a similar prognosis to those with small amounts of residual invasive carcinoma. Moreover, residual disease in lymph nodes was associated with a poor prognosis.

Evaluation by intrinsic subtype showed long-term survival benefit associated with pCR for triple-negative and ER−, HER2+ subgroups. In contrast, pCR in luminal A and luminal B subgroups did not have a prognostic impact, perhaps because these patients are more endocrine sensitive and receive extended adjuvant endocrine therapy.

This large scale analysis helps clarify the results from multiple studies with differing endpoints and sometimes varying results. It should prompt a consensus definition of pCR for subsequent neoadjuvant trials, so there is consistency in reporting. It also sets the stage for a new round of investigations into the molecular variations in the patients without pCR, including those with in situ disease only, to explain the prognostic effect of residual disease.

SUMMARY

OncologySTAT Editorial Team

Consistent with a rudimentary understanding of tumor biology, it has been hypothesized that a pathologic complete response (pCR) following neoadjuvant therapy for breast cancer would lead to improved long-term cancer-related outcomes, such as longer disease-free survival (DFS) and overall survival (OS). However, studies to date have been inconsistent in demonstrating an association between pCR and longer survival. Part of the issue has been the lack of a consensus definition for pCR—that is, does a response seen in the breast alone qualify as a pCR, or should all examined nodes also be clear of disease? The meaning of invasive vs in situ residual disease has also been variably interpreted. Furthermore, the prognostic impact of pCR may differ among histologic subtypes of breast cancer, as patients with luminal A tumors are known to have lower pCR rates but more favorable long-term outcomes than those with triple-negative tumors, where the converse has been true.

The authors of the current study performed a pooled analysis of data from seven breast cancer neoadjuvant therapy trials (GeparDuo, GeparTrio pilot and main study, GeparQuattro, AGO-1, PREPARE, and TECHNO), for a total of 6377 patients. All patients received at least one cycle of anthracycline/taxane therapy, and 622 patients from the GeparQuattro and TECHNO studies also received trastuzumab.

With a median follow-up of 46.3 months and almost 23,000 patient-years of observation, a total of 1466 (23%) relapses and 775 (12.2%) deaths had been seen. Median tumor size was 4.0 cm. There were 5618 (88%) patients with operable disease, while 759 (12%) had locally advanced disease. A total of 3771 (60.4%) tumors stained ER-positive, and 3235 (51.8%) were PR-positive. Due to inclusion of data from before 2001, 1990 (31.2%) patients did not have HER2 measurements.

The authors compared various definitions of pCR, of which ypT0 ypN0 (seen in 15% of patients and excluding any residual invasive, in situ, or nodal disease) produced the best DFS (even compared with ypTis ypN0 tumors with residual in situ disease in the breast). The worst DFS and OS were seen in patients with ypT0/is ypN+ tumors, where in situ disease could remain in the breast along with nodal positivity. Patients with ypT1mic tumors also fared worse than those with ypT0.

Histologic breast regression score (RS) did correlate significantly (P <.001) with OS and DFS. There was not a significant difference in outcome between patients with non-invasive residuals and those with focal-invasive residuals or those with minimal/no regression. Tumor stage after therapy (ypT) was significantly associated with prognosis, with patients with ypT3 and ypT4 tumors having the worst outcomes. Similar results were seen with persistent nodal positivity, with patients with ypN2 tumors having a median DFS of 70 months vs 30 months for patients with ypN3 tumors.

Using a pCR definition of ypT0 ypN0, pCR (vs no pCR) predicted overall for a more favorable outcome, independent of age, tumor size, nodal status, or HER2 status. Histology and immunohistochemistry clearly affected the meaningfulness of pCR, insofar as pCR did not predict for DFS or OS in patients with lobular type, grade 1, or ER/PR-positive tumors. In other words, pCR did not predict for improved outcome in patients with tumor types with lower proliferation.

Prognosis by subtype was also examined with and without pCR. In the absence of pCR, survival patterns were similar to those observed in patients receiving chemotherapy in the adjuvant setting—that is, better outcomes in patients with luminal vs HER2-positive tumors. However, among patients achieving pCR, prognosis did not differ among subtypes.

Interestingly, pCR was not associated with improved outcome in the luminal B/HER2-positive group, where pCR rates were low irrespective of exposure to trastuzumab.

Ultimately, the authors concluded that, with a large sample size and fair numerical representation of histopathologic subpopulations, pCR was not associated with prognosis in patients with slowly proliferating tumors but could differentiate good vs poor prognosis in those with highly proliferative tumors, such as HER2-positive (non-luminal). Subgroup ypT0 ypN0 was deemed the best definition of pCR.

Hopeful