Immune checkpoint inhibitors as neoadjuvant therapy in early triple-negative breast cancer: A systematic review and meta-analysis
Niti Mittal1, Surjit Singh2, Rakesh Mittal1, Jyoti Kaushal1, Vivek Kaushal3
1 Department of Pharmacology, Postgraduate Institute of Medical Sciences, Rohtak, Haryana, India
2 Department of Pharmacology, All India Institute of Medical Sciences, Jodhpur, Rajasthan, India
3 Department of Radiation Oncology, Postgraduate Institute of Medical Sciences, Rohtak, Haryana, India
Correspondence Address:
Niti Mittal
Department of Pharmacology, Postgraduate Institute of Medical Sciences, Rohtak, Haryana
India
Source of Support: None, Conflict of Interest: None
CheckDOI: 10.4103/jcrt.jcrt_1867_20
Context: Immune checkpoint inhibitors combined with chemotherapy are being evaluated in neoadjuvant settings in early triple-negative breast cancer (TNBC).
Aim: To evaluate efficacy and safety of checkpoint inhibitors in early TNBC.
Methods: Electronic search was done using PubMed, EMBASE, Google Scholar, Cochrane Central Register of Controlled Trials and clinicaltrials.gov to identify relevant articles till October 31, 2020. Clinical trials evaluating checkpoint inhibitors as neoadjuvant therapy in early-stage TNBC were included. Outcomes assessed included pathologic complete response (pCR), event-free survival (EFS), and safety.
Statistical Analysis Used: Meta-analysis was conducted using Cochrane review manager (RevMan) version 5.4. Randomized controlled trials (RCTs) were assessed for quality using Cochrane Collaboration risk of the bias assessment tool, version 2.0 (ROB-2). GRADE analysis was done to assess the overall quality of evidence for all outcomes.
Results: Out of 116 studies screened, 5 RCTs were included in meta-analysis. Compared to control group, programmed death-1 (PD-1)/programmed death-ligand 1 (PDL-1) inhibitor group was associated with significant increase in rate of pCR (odd ratio [OR] =1.71 [1.38–2.11]; P < 0.00001) and EFS (1.77 [1.21–2.60]; P = 0.003). There was a significant increase in risk of serious adverse events (risk ratio [RR] =1.53 [1.28-1.83]; P < 0.00001), adverse events of special interest (AESI) of any grade (RR: 1.5 [1.34–1.69], P < 0.00001) and grade 3 or higher AESI (RR: 2.8 [1.87–4.19], P < 0.00001) with PD-1/PDL-1 inhibitors compared to control.
Conclusions: PD-1/PDL-1 inhibitors in combination with neoadjuvant chemotherapy for early TNBC show significant improvement in pCR irrespective of PDL-1 status and cancer stage.
Keywords: Atezolizumab, checkpoint inhibitors, durvalumab, pembrolizumab, triple-negative breast cancer
Triple-negative breast cancer (TNBC) refers to a subtype of breast cancer that is characterized by the absence of expression of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor 2 receptor (HER2). TNBC accounts for 10%–20% of all breast cancer cases.[1] The hallmark of this particular subtype of breast cancer is its aggressive nature, metastasis at an earlier stage, visceral metastasis, and worse prognosis compared to other types of breast cancer. Furthermore, lack of expression of hormone and HER2 receptors in TNBC renders pharmacological treatment with hormone-related endocrine therapies futile.[2] The cornerstone of therapy in TNBC has been platinum-based cytotoxic chemotherapy until recent times. However, with the advancement in molecular classification and genome sequencing, newer potential molecular targets in TNBC are being identified and explored.[2] Few molecularly targeted approaches to TNBC include immune check point inhibitors, poly (ADP-ribose) polymerase enzyme (PARP) inhibitors, antibody-drug conjugates, cyclin-dependent kinase 4/6 (CDK 4/6) inhibitors, and multikinase inhibitors.[3]
Immune checkpoints are the molecules present on certain immune cells which on binding with specific ligands on target cells prevent the T-cells from attacking these cells, thereby acting as “off-switch” for immune response. This mechanism prevents the immune system from attacking normal cells against foreign cells, hence limiting autoimmunity.[4] Examples of immune checkpoints present on T cells are programmed death-1 (PD-1) and cytotoxic T-lymphocyte-associated protein 4 (CTLA4).[4] Cancer cells in certain solid malignancies including melanoma, cancers of the lung, breast, bladder, colon, liver, head, and neck are known to express ligands for these checkpoints (e.g., programmed death-ligand 1 (PDL-1)) thereby evading the destructive antitumor immune response. Thus, immune checkpoint inhibitors, i.e., drugs targeting the immune checkpoints or their specific ligands hold promise in the treatment of such cancers.[5],[6],[7],[8] Various checkpoint inhibitors being explored for their potential as targeted immunotherapies include PD-1 inhibitors (pembrolizumab, nivolumab, cemiplimab), PD-L1 inhibitors (atezolizumab, avelumab, durvalumab), and CTLA4 inhibitors (ipilimumab, tremelimumab).[7],[8]
As per literature reports, 20%–30% of TNBC express PD-L1 mainly associated with tumor-infiltrating lymphocytes,[9] which supports the potential role of PD-1/PDL-1 inhibitors in this subtype of breast cancer. Hence, many researchers are trying to explore the therapeutic potential of agents inhibiting PD-1/PDL-1 in TNBC under various settings like early or advanced/metastatic TNBC, first or second/higher line of treatment, as monotherapy or combination therapy, etc. In fact, based on positive results in IMpassion 130 trial,[10] USFDA granted accelerated approval (on March 08, 2019) to atezolizumab (PDL-1 inhibitor) in combination with nab-paclitaxel for the treatment of locally advanced or metastatic TNBC.[11] However, none of the targeted therapies has gained a strong foothold to be approved in early-stage TNBC (stage I–III) in which the mainstay of treatment remains standard neoadjuvant and adjuvant chemotherapy. The prognostic benefit offered by chemotherapy in early TNBC is not phenomenal as is evident from literature reports showing 70% rate of 5-year metastasis-free survival and 30%–40% incidence of distant metastasis and cancer-related mortality.[1],[12] Hence, the adoption of alternative approaches in order to improve long-term clinical outcomes in early-stage TNBC is the need of the hour. In this direction, many clinical trials are evaluating the potential therapeutic benefit of combining checkpoint inhibitors with neoadjuvant chemotherapy (NACT) in early-stage TNBC although reports on their efficacy and safety have been inconsistent across different studies and with different agents.
Keeping this in mind, we planned to conduct a systematic review and meta-analysis to evaluate the efficacy and safety of the combination of checkpoint inhibitors with NACT compared to NACT given alone in patients having early-stage TNBC.
> MethodsProtocol and registration
This systematic review was conducted according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement[13] and guidelines of the Cochrane Handbook for Systematic Reviews of Interventions (version 6.1).[14] The protocol was registered in the PROSPERO (International Prospective Register of Systematic Reviews) database (CRD42020217831).
Criteria for study inclusion
Inclusion criteria were defined as clinical trials (phases I/II/III) evaluating the role of immune checkpoint inhibitors as neoadjuvant therapy in patients with early-stage TNBC. Exclusion criteria were quasi-randomized or cluster-randomized trials, and trials with incomplete efficacy and safety data which cannot be extracted even with the best possible efforts.
Search strategy and study selection
The electronic search was done using PubMed, EMBASE, Google Scholar, Cochrane Central Register of Controlled Trials, and clinicaltrials.gov for all relevant articles until October 31, 2020. Bibliographic searching was conducted for manual identification of further studies not indexed in the above databases. Gray searching was carried out to identify relevant published conference abstracts from the following organizations: American Society of Clinical Oncology, American Society of Clinical Oncology Breast Cancer Symposium, European Society of Medical Oncology (published in the Annals of Oncology), and European Council for Clinical Oncology (published in the European Journal of Cancer). Articles published in English were included in the review.
Search strategy comprised the following search terms along with their associated medical subject headings: “breast neoplasms,” “triple-negative breast cancer,” “check point inhibitors,” “programmed death inhibitors,” “programmed death-ligand inhibitors,” “pembrolizumab,” “atezolizumab,” “durvalumab,” “nivolumab,” “avelumab,” “ipilimumab.” Two independent researchers (NM and SS) assessed the titles and abstracts retrieved by electronic searching for potential eligibility followed by quality assessment of eligible studies. Disagreement, if any was resolved by consensus with other review authors [Supplementary File 1].
Data extraction
Two independent review authors (NM and SS) used pre-tested structured data collection form to extract information regarding general study information, interventions and regimens, subject characteristics, efficacy, and safety outcomes reported.
Study outcomes
The primary efficacy outcome was pathologic complete response (pCR, defined as the absence of invasive tumor in the breast and regional lymph nodes at the time of surgery). For pCR, subgroup analysis was planned for PDL-1 status (PDL-1 positive, i. e., PDL-1 combined positive score ≥1 and PDL-1 negative), cancer stage (I-II and III-IV), lymph nodal status (positive and negative), and ECOG performance status score (0 and 1).
Secondary efficacy endpoint included event-free survival (EFS, defined as the number of patients who were alive without disease progression/local or distant recurrence/second primary tumor). Safety endpoints included all serious adverse events (SAE; classified according to the WHO toxicity criteria or National Cancer Institute's Common Terminology Criteria for Adverse Events) and adverse events of special interest (AESI).
Quality assessment
Cochrane collaboration risk of bias (ROB) assessment tool (version 2) was used to assess the methodological quality of included trials by two independent reviewers[15] Robvis (visualization tool)[16] was used to synthesize the plots for ROB.
Data synthesis and summary measures
Odds ratios (OR) and risk ratios (RR) 95% confidence intervals (CI) were used to summarize the dichotomous data. Meta-analysis was conducted using Review Manager 5 (RevMan) Version 5.4. I2 was used to assess heterogeneity with values of 25%, 50%, and 75% indicating low, medium, and large heterogeneity, respectively.[17],[18] The results of the fixed effect model using the generic inverse variance approach were presented. P <0.05 was considered significant.
Assessment of quality of evidence-GRADE Pro analysis
GRADE pro-guideline development tool software was used to assess the overall quality of evidence for each outcome.[19],[20] Optimal information size was calculated as 275 patients in each group.
> ResultsStudy selection
[Figure 1] depicts the study selection process.
Study characteristics
[Table 1] gives the characteristics of included studies. Of the 6 studies, there were 5 randomized controlled trials (RCTs)[21],[22],[23],[24],[25] and 1 single-group trial.[26] All the RCTs had combination of check point inhibitors with standard NACT (sequential use of a taxane and anthracycline plus cyclophosphamide) in the intervention arm. Various check point inhibitors evaluated included PD-1 inhibitors (pembrolizumab) and PD-L1 inhibitors (atezolizumab and durvalumab). From the 5 RCTs included in meta-analysis, total number of evaluable patients included 822 and 674 patients in intervention and control arms, respectively.
Table 1: Characteristics of clinical trials evaluating programmed death-1/programmed death ligand 1 inhibitor as neoadjuvant therapy in early stage triple-negative breast cancerQuality of included studies
The overall ROB was recorded as low in three RCTs. Risk of bias was recorded as high for Nanda et al.[21] (due to deviations from intended interventions) and Gianni et al.[24] (due to deviations from intended interventions and bias in missing outcome data) [Figure 2]. The results of the study by Gianni et al. were obtained from an interim unpublished report.[27] Summary weighted ROB is represented in [Supplementary Figure 1].
Figure 2: RoB-2: RoB in randomized clinical trials evaluating PD-1/PDL-1 inhibitors as neoadjuvant therapy in early TNBC. PD-1 = Programmed death-1, PDL-1 = Programmed death ligand 1, RoB = Risk of bias, TNBC = Triple negative breast cancerEfficacy outcomes
Pathological complete response
In pooled analysis, statistically significant improvement in pCR rate in PD-1/PDL-1 inhibitor group was observed compared to control (OR = 1.71 [95% CI: 1.38–2.11]; P < 0.00001, I2 = 61%) [Figure 3].
Figure 3: Pooled analysis of differences in percentages of patients with Pathological complete responseThe improvement in pCR seen in the PD-1/PDL-1 inhibitor group was consistent across subgroups except for 2 viz., nodal status negative and ECOG performance status score 1 for which statistical significance could not be attained [Figure 4].
Figure 4: Subgroup analysis of differences in percentages of patients with Pathological complete response.Event free survival
EFS rate was significantly greater in PD-1/PDL-1 inhibitor group compared to control (OR = 1.77 [95% CI: 1.21–2.60]; P = 0.003) [Figure 5]. Result for EFS was extracted from 2 trials with 812 and 470 patients in pembrolizumab and control groups, respectively.
Safety outcomes
Serious adverse events
There was a significant increase in the risk of SAEs with PD-1/PDL-1 inhibitors compared to control (RR = 1.53 [95% CI: 1.28–1.83], P < 0.00001); data extracted from 3 RCTs with 1037 and 638 patients in intervention and control groups, respectively [Figure 6]a. Most common SAE reported was neutropenia although the increase in risk with PD-1/PDL-1 inhibitors was not significant compared to control (RR: 1.02 [0.89–1.18], P = 0.74).
Figure 6: (a) Serious adverse events (PD-1/PDL-1 inhibitors vs. standard). (b) Adverse events of special interest (any grade) (PD-1/PDL-1 inhibitors vs. standard). PD-1 = Programmed death-1, PDL-1 = Programmed death ligand 1Grade 3 or higher treatment related AEs (TRAEs) were not significantly increased in treatment group compared to control (RR = 1.06 [95% CI: 0.99–1.14], P = 0.08) [Supplementary Figure 2].
Adverse events of special interest
Statistically significant increased risk of AESI of any grade (RR: 1.50 [1.34–1.69], P < 0.00001) [Figure 6]b and grade 3 or higher AESI (RR: 2.80 [1.87–4.19], P < 0.00001) [Supplementary Figure 3] was noted with PD-1/PDL-1 inhibitors compared to control. Most common AESI reported were infusion reactions (RR: 1.63 [1.19–2.22], P = 0.002) [Figure 7]a and thyroid dysfunction (RR: 3.25 [2.28–4.62], P < 0.00001) [Figure 7]b. Results for this outcome were extracted from 4 RCTs with 1176 and 777 patients in PD-1/PDL-1 inhibitor and control arms, respectively.
Figure 7: (a) Infusion related reactions (PD-1/PDL-1 inhibitors vs. standard). (b) Thyroid dysfunction (PD-1/PDL-1 inhibitors vs. standard). PD-1 = Programmed death-1, PDL-1 = Programmed death ligand 1Publication bias
The funnel plot of pCR (PD-1/PDL-1 inhibitors) appeared symmetrical around effect estimate [Figure 8]. Egger's regression test for funnel plot asymmetry showed low publication bias for pathological complete response (t = 0.9670, P = 0.4049), AESI (t = −1.5094, P = 0.2703) and thyroid dysfunction (t = 0.2510, P = 0.8252). As there were fewer than five studies, we did not analyze publication bias with the help of funnel plot for EFS.
Figure 8: Funnel plot depicting publication bias for studies included in the reviewGRADE analysis of the primary and secondary outcomes
GRADE recommendation for pCR (PD-1/PDL-1 status), as well as EFS, was “high.” This was due to overall low ROB, less heterogeneity, and high precision. For pCR (PD-1/PDL-1 inhibitors), the GRADE recommendation was “Moderate” quality evidence. For SAE, AESI and thyroid dysfunction, GRADE was recommended as “MODERATE” due to high heterogeneity [Table 2].
Table 2: Grading of recommendations, assessment, development and evaluations recommendation for outcomes of systematic review of programmed death/programmed death ligand 1 inhibitors as Neo adjuvant therapy in early triple negative breast cancer > DiscussionTNBC continues to pose a therapeutic challenge because of its grave prognosis and lack of standard treatment. Many molecularly targeted approaches are being tested for their potential role in this subtype of breast cancer. The approval of few targeted therapies including PARP inhibitors (olaparib and talazoparib)[3] and PDL-1 inhibitor (atezolizumab)[11] for treatment of advanced or metastatic TNBC fueled the burgeoning of trials to explore their therapeutic potential in early-stage disease as well. Till date, the use of NACT is a standard practice in early-stage breast cancer.[28] In this review, we planned to test whether combining checkpoint inhibitors to NACT offers therapeutic advantage in terms of efficacy and safety over standard NACT given alone in patients with early (stage I-III) TNBC.
Historically, attainment of pathological complete response (pCR) in TNBC has demonstrated effective correlation with EFS rates.[29] Thus, pCR is accepted as a surrogate endpoint for long-term clinical outcomes in early-stage cancers. In 2014, the FDA released guidance on pCR as an endpoint for accelerated approval of neoadjuvant treatment in TNBC and similar breast cancers.[30] Hence, we selected pCR as the primary efficacy endpoint in our meta-analysis. We observed a statistically significant improvement in the rate of pCR with the combination of PD-1/PDL-1 inhibitor and NACT compared to standard NACT alone. All the PD-1/PDL-1 inhibitors showed improvement in pCR rate although statistical significance could not be achieved with durvalumab. In the durvalumab study,[25] a 2-week window period before nab-paclitaxel was included during which patients were administered durvalumab or placebo. It was observed that patients who received durvalumab in this window period had significant improvement in pCR rate (OR: 2.22 [95% CI: 1.06–4.64]; P = 0.03) which was not seen in the other arm (OR: 0.61 [95% CI: 0.21-1.75]; P = 0.36). Hence, priming with durvaluamb seems to play an important role in improving the rate of pCR which however needs to be confirmed in future clinical trials.
The improvement in pCR rate was consistent across PDL-1 positive and negative subgroups. This observation is in contrast to the reported benefit of atezolizumab on median overall survival and median progression-free survival confined specifically to PDL-1 positive patients in IMpassion 130 trial[10] which formed the basis for restricted regulatory approval of atezolizumab in PDL-1 positive metastatic TNBC. A possible explanation for such contrasting finding is the difference in disease stage viz. early vs advanced/metastatic which may act as an independent predictor to response to immune checkpoint inhibition. It has been hypothesized that the tumor immune microenvironment in early-stage TNBC is more robust to the anti-tumor response of immunotherapy compared to the metastatic stage so that PDL-1 status plays a minor role in determining response to immunotherapy in early-stage cancer.[31] Hence, unlike in metastatic disease, the use of PD-1/PDL-1 inhibitors as neoadjuvant therapy in early TNBC is deemed to be more broadly applicable.
Significant improvement in EFS with PD-1/PDL-1 inhibitors compared to control as reported in our analysis was consistent with results of pCR. The results on EFS were extracted from 2 trials with pembrolizumab,[21],[22] out of these one was the first interim analysis of ongoing trial with insufficient follow-up duration to report longer EFS.[22] Two trials included in this review[23],[25] were not powered for EFS. Gianni et al.[24] have included 5 year EFS as the primary endpoint, however, the first interim report included in this review has not reported data with respect to EFS. Hence, the precise prognostic impact of PD-1/PDL-1 inhibitors as neoadjuvant therapy in early TNBC needs to be confirmed with data on survival analysis in future long-duration trials adequately powered for this outcome.
In our meta-analysis, the fact that significant improvement in pCR rate in the intervention arm (PD-1/PDL-1 inhibitors combined with NACT) was observed across various stages of TNBC (stage I to IV) and with lymph node involvement favors the use of these agents, especially in patients with advanced stages and high tumor load which are less responsive to chemotherapy alone as compared to earlier stage tumors (stage I to II). The results of our meta-analysis are in congruence with a review by Zhao et al. suggesting the role of PD-1/PDL-1 inhibitors as a potential therapeutic strategy in patients with TNBC.[33]
Serious adverse events (neutropenia, anemia, and febrile neutropenia) were reported more in the intervention group, however, they were consistent with the safety profile of co-administered cytotoxic chemotherapy. Moreover, the risk of neutropenia in the PD-1/PDL-1 inhibitor group was not significantly higher than control. PD-1/PDL-1inhibitors were not associated with the significant increase in the risk of grade 3 or higher TRAEs.
Potentially, immune-mediated AESI associated with immune checkpoint inhibitors as a class includes hypo/hyperthyroidism, infusion reactions, pneumonitis, rash, pruritis, hepatitis, and adrenal insufficiency.[32] Significant increases in the risk of AESI particularly infusion reactions and thyroid dysfunction were observed with PD inhibitors compared to control. Although the majority of these were of lower grade and manageable but being a new class of toxicity, attention needs to be paid for their appropriate identification and management.
Limitations and strengths
A major limitation of our meta-analysis is the extraction of data from the limited number of available randomized controlled trials which raises concerns regarding reliable subgroup and sensitivity analyses.
GRADE analysis is one of the major strengths of our review. Due to “Moderate” to “High” GRADE evidence for critical efficacy and safety outcomes, the results of these were interpreted and used for drawing conclusions.
Quality of evidence-GRADE Pro analysis
The overall quality of systematic review is “Moderate” as the critical outcomes like pCR (PD-1/PDL-1 status) and EFS (in intent to treat population) had “High” quality of evidence while pCR (PD-1/PDL-1 inhibitors) had “Moderate” quality. Safety outcomes viz., SAEs, AESI, and thyroid dysfunction have “Moderate” quality evidence. Therefore, overall GRADE was recommended as “Moderate” quality evidence implying the potential of future research to have a bearing on the conclusions of this review.
> ConclusionsSignificantly favorable results in efficacy outcomes viz., pathological response rate and EFS with PD-1/PDL-1 inhibitors when given in combination with NACT in early TNBC was evident irrespective of disease stage and PDL-1 expression status. Moderate-to-high-quality evidence generated for these parameters carries conviction for their use as neoadjuvant therapy in early TNBC, although data on survival needs to be accrued to further substantiate their prognostic benefit in this condition. Safety concerns like AESI with “Moderate” quality GRADE evidence however warrant judicious and monitored use of these agents. With current evidence, it can be concluded that combined use of PD-1/PDL-1 inhibitors with NACT in early TNBC offers a therapeutic advantage irrespective of PDL-1 status and disease stage which broadens their applicability and favorable use, particularly in patients with advanced tumor stages who are less responsive to chemotherapy alone.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
Supplementary File 1: Search strategy for PubMed
> References
留言 (0)