As well as HR and HER2 status, some studies have indicated that the histological subtype of the breast cancer also plays an important role in predicting the response to adjuvant chemotherapy and/or neoadjuvant chemotherapy (NAC)[16-19]. In 2005, Cristofanilli et al[16,17] reported that ILC is characterized by lower pathologic complete response (pCR) rates to NAC but better long-term outcomes compared to IDC. In 2007, Katz et al[18] reviewed randomized trials of NAC and noted that the pCR rate was 1.7% in ILC and 11.6% in IDC (no special type). In 2010, in the era of tailored therapy for individual patients, Purushotham et al[19] documented that we would no longer routinely recommend NAC in patients with ER-positive, HER2-negative, classical type ILC.
However, though it is generally admitted that ILC, especially HR-positive, HER2-negative ILC, responds poorly to chemotherapy, currently available data do not unanimously support these assumptions. In 2012, Lips et al[9] reported a similar pCR rate in both ER-positive, HER2-negative IDC and ER-positive, HER2-negative ILC patients (4.2% and 4.3%, respectively). In 2014, Guiu et al[20] reported that in multivariate analysis, histology of ILC was not an independent negative predictive factor of pCR in seven[21-27] of nine studies[28,29,21-27].
Thus, we could not draw a conclusion that ILC or even HR-positive, HER2-negativeis ILC is an independent predictor of poor response to adjuvant chemotherapy and/or NAC. In fact, minority of past and current studies take lobular histology into account in pretreatment stratification or subgroup analysis. Consequently, findings of these studies limit our ability to indicate whether patients with IDC or ILC should be managed with similar or different treatments. At present, the National Comprehensive Cancer Network (NCCN) and the St Gallen International Expert Consensus guidelines for systemic therapy decisions are almostly derived from studies based on IDC. Neither of these two guidelines consider histologic subtype as a factor for determining systemic therapy decisions. Making systemic therapy decisions for patients with ILC is thus challenging for the oncology community. It is unlikely that a future randomized clinical trial (RCT) concerning this subject will be accomplished. There is lack of stronger evidence support, this may be why our guidelines still do not distinguish ILC from IDC for treatment allocation or classification therapy.
In this study by using SEER database, we firstly compared the cohort characteristics between the included HR-positive, HER2-negative, pT1b-c/N0-1/M0 ILC patients with and without adjuvant chemotherapy, in both original and propensity score matched sample, respectively. Secondly, OS and BCSM analyses between chemotherapy and control groups were made, before or after PSM, respectively. Thirdly, to adjust the potential confounding factors to chemotherapy, the multivariate Cox regression analyses were performed for overall survival and BCSM, in both original and matched sample, respectively. Our data demonstrate that patients with HR-positive, HER2-negative, pT1b-c/N0-1/M0 ILC could not derive survival benefit from the adjuvant chemotherapy shown in Fig.2, Fig.3 and Table 3, neither for OS nor for BCSM. In both original and propensity score matched sample, ILC patients with adjuvant endocrine therapy and chemotherapy had a worse BCSM than ILC patients with adjuvant endocrine therapy alone. This finding is almost certainly secondary to selection bias and not cause and effect of adjuvant chemotherapy.
Histological grading is an important part of breast cancer classification, and is performed using the Nottingham histological grading system. However, it has been controversial as to the relevance of this system for ILC, since tubule formation is rare (except in the tubulo-lobular variant)[30]. With limited nuclear pleomorphism and sparse mitotic count, ILC (including variants) is often characterized by lower histologic grade compared to IDC[31]. In both our original and matched samples, almost or more than ninety percentages of ILC were histologic grade 1-2 (Table 1 and Table 2). Consequently, a therapeutic dilemma can occur in the event of the relative resistance of ILC to conventional chemotherapeutic agents[32,33]. Moreover, lack of E-cadherin protein expression in ILC is distinctive from IDC[34]. It has been hypothesized that the lack of chemosensitivity of ILC is explained by the inactivation of E-cadherin in ILC. Loss of E-cadherin protein is thought to increase of epithelial to mesenchymal transition (EMT), which in turn become more resistant to chemotherapy[35]. Accordingly, lower histologic grade and deficiency of E-cadherin of ILC both supported our results.
It has been demonstrated that ILC and IDC have distinct genomic, transcriptomic and expression profiles[36]. Recent major advances in genome-wide transcription analyses, comparative genomic hybridization (CGH) and genomic tests further acknowledged the natural history and also the heterogeneity of ILC[37]. It has been suggested that the genomic signatures could be used to assist systemic therapy decisions for patients with ILC, and especially the decision of adding chemotherapy to hormonal therapy[38]. For instance, mutations in exon 9 of the PIK3CA gene have previously been reported more frequent in ILC than in IDC[39-41]. These mutations increase kinase activity, confer increased resistance to paclitaxel and are associated with metastatic capability[42,43]. Intriguingly, loss of E-cadherin of ILC has been also associated with many genetic and molecular alterations including the inactivation of the CDH1 gene at 16q22 by mutation, loss of heterozygosity, or CDH1 promoter methylation, which finally lead to the poor response to cytotoxic chemotherapy[4,3,44].
Oncotype Dx Breast Cancer Assay is a 21-gene assay used in estrogen receptor (ER)-positive breast cancer to predict benefit from chemotherapy[45,46]. In 2015, Conlon et al[47] reported that Oncotype Dx recurrence score (RS) currently plays a clinically useful role in the management of ILC, which may prevent the over-treatment of adjuvant chemotherapy. In 2017, Kizy et al[48] reported that patients with ER-positive ILC, 8% were in the high-risk and 72% were in the intermediate-risk groups as per the trial assigning individualized options for treatment (TAILORx) RS cutoffs. Adjuvant chemotherapy did not seem to confer a survival benefit for either the intermediate- or the high-risk cohorts[48].
Some limitations of our study have to be considered, thus we ought to be caution about our results. Our SEER database does not include information regarding the ILC and its variants, the loss of E-cadherin, the exact ER and PR and Ki67 expression, the 21-gene assay, the administration of chemotherapy and endocrine therapy, ect. Additionally, we should exclude all cases where breast cancer has only been reported by death certificate or autopsy. Thirdly, there is indeed an important deficiency is that the chemotherapy record in SEER database is classified as "No/Unknown" and "Yes". Although we obtained data of 1785 patients with definite chemotherapy from SEER database, we don't know whether the patients recorded as "No/Unknown" actually received chemotherapy. All these confounding factors may have affected our results. For example, the most recent 2012 WHO classification of breast cancer distinguishes the ILC and its variants: classic, solid, alveolar, pleomorphic, tubulo-lobular, and mixed variant [1]. Lack of E-cadherin is observed in all histological ILC variants, except for tubulo-lobular variant (tubulo-lobular carcinoma, TLC). Pleomorphic variant (pleomorphic invasive lobular carcinoma, PILC) shares many additional genomic changes with classic ILC such as TP53 stabilization, amplifications of MYC, MDM2, HER2/TOP2A and 20q13[49].
Our study is subject to some methodologic limitations too, which will lead to inevitable bias. The present study is a retrospective cohort study, however, not a RCT. The patient demographics and tumor characteristics are not totally consistent between the included ILC patients with and without adjuvant chemotherapy, even though after PSM analysis. Furthermore, the PSM analysis is also limited by the lack of adjustment for the cointervention of surgery therapy or radiation therapy, which demotivates our study due to reduce the sample sizes or event rates.
Nevertheless, until now, it is not clear whether there is a difference ineffectiveness between chemotherapy regimens administered to patients with ILC. Therefore, we suggest that further research on the type of chemotherapy administered to patients with ILC should be carried out. Moreover, evaluation of the response of ILC patients to endocrine therapy is an emerging direction of clinical breast cancer research[50]. It was reported that the magnitude of benefit of adjuvant letrozole is greater for patients diagnosed with ILC compared to those with IDC[51]. In fact, it may be time for the oncologists to consider a prospective RCT to evaluate the role of NAC versus neoadjuvant endocrine therapy in ILC patients [18]. Additionally, whether CDK4/6 inhibitor is more effective for HR-positive, HER2-negative ILC than for HR-positive, HER2-negative IDC is worth to study. Finally, we advise to the oncologists that ILC and its variants should be studied, with further efforts made to develop more individualized treatment for them and to identify potential mechanisms of their biological inferiority and superiority, respectively[52,53].