In this retrospective analysis of blood cell counts and ratios, as easily detectable pro-inflammatory parameters, MLR and NLR at baseline were shown as independent predictive marker usable for therapy guidance in mBC patients to receive CDK4/6i plus ET. CDK4/6i treated patients with high MLR at baseline were significantly more likely to achieve no clinical benefit, indicating progressive disease within the first six months of therapy. Moreover, the prognostic value of MLR and NLR at baseline with regard to having a shorter OS was shown for patients receiving CDK4/6i as first line therapy after diagnosis of metastasis.
By comparison of matched samples at baseline and after four weeks of therapy, a significant decrease of neutrophils, monocyctes, platelets, leukocytes, lymphocytes, eosinophils as well as MLR and NLR under CDK4/6i was detected. In contrast, MCV increased in the majority of patients under therapy.
While decreasing PLR and increasing MCV within the first cycle of CDK4/6i correlated with shorter PFS, decreasing MLR within the first cycle was only correlated with a shorter PFS in the 1L CDK4/6i treated patients. These early-on-treatment assessments could be used as monitoring marker for therapy success.
Standard-of-care first-line therapy for patients with HR+/HER2- mBC without visceral crisis is CDK4/6i combined with ET, resulting in substantial PFS and OS benefits as well as maintained quality of life compared to ET alone7,8. HR+/HER2- mBC patients with visceral crisis receive chemotherapy as standard-of-care7. However, it is to question whether some patients without visceral crisis would benefit from chemotherapy more than from CDK4/6i plus ET. Although a meta-analysis revealed that no chemotherapy regimen showed increased PFS compared to CDK4/6i plus ET33, the Pearl study did not show superiority in PFS of Palbociclib+Fulvestrant vs. Xeloda34. Presented as an abstract, the Right Choice study further postulated that first-line Ribociclib plus ET increased PFS from 12.3months to 24.0months compared to chemotherapy in a patient cohort that included more than 50% of patients with visceral crisis35. These two studies question the standard-of-care and highlight the importance of individual factors that mediate the outcome under CDK4/6i.
At the moment, research focusses on these kind of individual factors as predictive markers to indicate de novo resistance to CDK4/6i therapy36. Predictive markers would enable individualized therapy approaches towards longer PFS, spared side effects and increased quality of life.
One of the factors that might influence the outcome of CDK4/6i is the composition of the tumor microenvironment37, because CDK4/6i was shown to effect the immune system. CDK4/6i triggers anti-tumor immunity by different ways. CDK4/6i activates tumor cell expression of endogenous retroviral elements, thus, increasing intracellular levels of double-stranded RNA. This in turn stimulates production of type III interferons and enhances tumor antigen presentation28. Enhanced tumor antigen presentation leads to increased anti-tumor immunity due to detection and killing of tumor cells by cytotoxic T cells. Second, the cytotoxic T cells themselves are also effected by CDK6 inhibition as it leads to de-repression of NFAT family transcription factors and consequently, increased cytotoxic T cell recruitment and enhanced T cell activation29. Both effects lead to cytotoxic T cell-mediated clearance of tumor cells. Third, CDK4/6 inhibitors markedly suppress the proliferation of regulatory T cells (Tregs) by reduced activity of the E2F target, DNA methyltransferase28. Consequently, the number of pro-tumorigenic Tregs decrease. In summary, CDK4/6 inhibitors decrease Treg proliferation but increase tumor infiltration and activation of cyctotoxic T cells leading to an overall enhanced anti-tumor immunity. Fourth, CDK4/6i also leads to reduced stem cell and progenitor cell proliferation mediated by reduced Notch signaling30. A consequence of these CDK4/6i effects might be the long-term reduction in different blood cell populations and the common adverse events like neutropenia and leukopenia.
With the knowledge of the numerous effects of CDK4/6i therapy on tumor cells, Tregs, cytotoxic T cells, and even stem and progenitor cells, it is reasonable to assume that the pretreatment status of the tumor immunity may have predictive value. NLR, MLR and PLR are pro-inflammatory signatures representing peripheral blood surrogates of the tumor immunity.
Comparison of the descriptive statistics of all blood parameters at baseline in the entire HR+/HER2- mBC cohort to the healthy donor reference ranges according to Wakeman et al.32 showed the minimal and maximal values of the entire HR+/HER2- mBC cohort within the healthy reference range for lymphocyctes, eosinophils and basophils. In contrast, in the entire HR+/HER2- mBC cohort, minimal and/or maximal values for neutrophils, monocyctes, platelets, leukocytes and MCV were outside the healthy reference range. The mean values of all mentioned eight blood parameters of the entire HR+/HER2- mBC cohort at baseline were within the healthy reference range.
We identified high MLR at baseline as independent predictive factor for reduced PFS in the CDK4/6i cohort and decreasing MLR within the first four weeks under therapy was associated with poor PFS in the CDK4/6i 1L cohort. To our knowledge, MLR has not been examined as predictive marker for CDK4/6i by any other group before. It is further to notice that we used a similar cut-off as other groups that studied MLR in mBC patients including all BC subtypes (0.36 in this project; 0.3438 and 0.2839).
In a mouse model representing lung metastases in the BC setting, it was shown that circulating monocyctes were reduced in number under CDK4/6i, but an increase in monocyte invasion was detected40. The decrease in circulating monocyte number was detected in CDK4/6i patients in our project as well.
A meta-analysis revealed that the NLR cut-off values in 15 analyzed studies ranged from 1.9 to 5.041. The mean NLR value of 2.98, used in our entire cohort as cut-off, is in line with the majority of other studies that used 3.0 as cut-off41, further justified in one study as the optimal value to differentiate mBC patients with OS less or greater than 24 months38.
We identified high NLR at baseline as an independent predictive marker for shorter PFS in the CDK4/6i cohort. These results confirm the results shown recently as conference abstract, where high NLR at baseline was independently associated with lower PFS in 308 HR+/HER2- advanced BC patients receiving CDK4/6i therapy42. Similar results were demonstrated in a more stringent cohort of 126 1L CDK4/6i patients with a NLR cut-off of 2.5343. In a smaller cohort of 89 1L CDK4/6 treated HR+/HER2- mBC patients, high NLR (here defined>3.7) was not found to be significantly correlated with worse PFS in a meeting abstract44. The latter two studies highlight the importance of evaluating markers in a large stringent cohort and only the use of consistent cut-off values will lead to reproducible results transferable into clinical practise. However, the evidence of high NLR at baseline as predictive marker for PFS under CDK4/6i therapy accumulates and might, thus, be usable to identify patients with de novo resistance to CDK4/6i due to the pretreatment status of the tumor immunity detected by blood surrogates.
In our study, we included a control cohort only receiving ET and found no significant lower PFS in patients with high NLR at baseline. However, in the entire cohort, including CDK4/6i treated and only endocrine treated patients, univariate Cox regression showed a significant association of high NLR with worse PFS, questioning the specificity of NLR with regard to the given therapy as predictive marker for CDK4/6i. Similarly, the results of two studies analyzing eribulin treated mBC patients also suggested high NLR to be a predictive marker for shorter PFS in mBC receiving other treatment regimens than CDK4/6i22,45.
In this regard, it is to mention that clinical parameters influencing PFS should be integrated in the identification of a predictive marker, as done in this study by multivariate Cox regression analysis with clinical parameters also used in other CDK4/6i studies46,47. This is important because in a cohort of 263 mBC patients including all BC subtypes, a high NLR (defined as>2.32) was significantly associated with worse PFS in univariate Cox regression but not in multivariate Cox regression analysis48.
Interestingly, Kim et al. determined NLR after one treatment cycle with high NLR to predict a reduced PFS49,50. However, this strategy is not usable for therapy decision making, because a predictive marker has to be evaluated before therapy start.
In addition to the predictive value of NLR, we and others43 showed a significant correlation of high NLR at baseline with shorter OS in the 1L CDK4/6i cohort, demonstrating the prognostic value of NLR.
Decreasing PLR from baseline to four weeks of therapy correlated with a shorter PFS and OS in the CDK4/6i cohort identifying PLR dynamics as a potential monitoring and prognostic marker.
PLR, not PLR dynamics, correlated with worse OS in a huge cohort of 2374 BC patients, using a cut off value>30051. However, this group found high PLR not related to OS when analyzing only the luminal BC patients, which is in line with our results, since baseline PLR values in our luminal mBC patients before CDK4/6i showed no prognostic value.
Zattarin et al., described in a conference abstract that PLR at baseline and also after the first three treatment cycles related to worse PFS in 308 HR+/HER2- advanced BC patients receiving CDK4/6i42 while Weiner et al., presented in a meeting abstract a significant association between PLR at baseline and PFS using univariate und multivariate analysis in a more stringent cohort of 89 1L CDK4/6i treated CDK4/6 patients44. Both groups revealed a predictive value of PLR at baseline before CDK4/6i in mBC which we cannot confirm in our study.
Using an identical mean MCV value of 89.0fl at baseline as described before52, the rising MCV under CDK4/6i in our patients confirmed already published results52,55,56. In contrast to the results of the other workgroups, our resultsshowed a prolonged PFS in patients with increasing MCV even in multivariate Cox regression analysis.
Despite the search for a predictive marker assessable at baseline for therapy decision making, we evaluated the dynamics of the blood cell counts and ratio under the first CDK4/6i cycle. NLR, MLR as well as the number of neutrophils, monocyctes, platelets, leukocytes, lymphocytes and eosinophils decreased significantly from baseline to four weeks under therapy. These dynamics might be explained by the CDK4/6i induced cell-cycle arrest in hematopoietic cells30 but also by the increased recruitment of cyctotoxic T cells29 as well as the proliferation repression of regulatory T cells28 by CDK4/6i which we can only speculate but not proof in this study.
In addition to the monitoring value of MCV shift under CDK4/6i, we here present the monitoring value of the MLR shift from baseline to four weeks under therapy in the 1L CDK4/6 cohort, that might also be used to identify therapy success during treatment. Furthermore, we detected a prognostic value of baseline MLR, NLR and PLR in the 1L CDK4/6 cohort suitable to assess the outcome early and adjust therapy management and follow-up care.
Since this was a retrospective study, time points of blood cell count analysis varied in a number of cases for baseline and four-week analysis. Consequently, especially the short-term effects of one therapy free week after CDK4/6i therapy at the end of each cycle could have influenced the blood results. Despite evaluating different blood cell counts and ratios, detailed analysis of other inflammatory markers, eg. C-reactive protein, procalcitonin or other acute-phase proteins and Lactate dehydrogenase should also be taken into account in future studies. Unfortunately, these protein quantities were not available for our patients. Acute or chronic inflammatory diseases and/or cortison, novalgin or non-steroidal anti-inflammatory drug etc. intake was not documented for the included patients. Further prospective studies using fresh blood at the given time points would also allow for the quantification of B cells and T cells or the CD4+/CD8+cell ratio.
Although limited in sample size and not randomized for different treatment regimens, one advantage of our study is a control group to identify CDK4/6i specificity within our findings. To the best of our knowledge, this is the first study which included a control group to clearly identify CDK4/6i therapy specific predictive markers. One further advantage of this study are clearly defined sampling time points independent of dose reduction and drug holiday, which only take place after more than one therapy cycle.
We included patients that received CDK4/6i in the first line as well as in second or more lines, thus, the results of the latter group have to be interpreted with caution due to unknown effects of prior therapies on blood cell counts. The majority of patients (29/33) in the2L CDK4/6i received Palbociclib plus ET and only four patients in this cohort received Ribociclib plus ET, thus, results obtained for patients with2L CDK4/6i may be related more to Palbociclib then Ribociclib. However, our subgroup stratification clearly showed significances not only in the 1L CDK4/6i cohort, but also in the CDK4/6i cohort, consisting of 1L and2L CDK4/6i treated patients. Consequently, we here demonstrate a broader range of implications for the clinical setting as our results are not specific for the number of therapy lines applied before start of CDK4/6i therapy, but can be applied to the entire HR+/HER2- mBC population receiving CDK4/6i. As we differentiated the 1L CDK4/6i cohort from the2L cohort for the statistical analysis, reported 1L cohort results are as stringent as in other studies that only analyzed 1L CDK4/6i patients. The subgroup stratification into 1L and2L CDK4/6i recently gained relevance with the publication of the SONIA trial results that challenged the need of CDK4/6i in the 1L57.
Read more here:
Ratios of monocytes and neutrophils to lymphocytes in the blood ... - Nature.com
Recommendation and review posted by G. Smith
