Development of Predictive Models for Pathological Response Status in Breast Cancer after Neoadjuvant Therapy Based on Peripheral Blood Inflammatory Indexes

doi:10.21203/rs.3.rs-4917041/v1

Download PDF

Research Article

Development of Predictive Models for Pathological Response Status in Breast Cancer after Neoadjuvant Therapy Based on Peripheral Blood Inflammatory Indexes

https://doi.org/10.21203/rs.3.rs-4917041/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 12 Oct, 2024

Read the published version in BMC Women's Health →

You are reading this latest preprint version

Background.

Achieving a pathological complete response (pCR) after neoadjuvant therapy (NAT) is considered to be a critical factor for a favourable prognosis in breast cancer. However, discordant pathological complete response (DpCR), characterised by isolated responses in the breast or axillary, represents an intermediate pathological response category between no response and complete response. This study aims to investigate predictive factors and develop models based on peripheral blood inflammatory indexes to more accurately predict NAT outcomes.

Method.

A total of 789 eligible patients were enrolled in this retrospective study. The patients were randomized into training and validation cohort according to a 7:3 ratio. Lasso and uni/multivariate logistic regression analysis were applied to identify the predictor variables. Two Nomograms combining clinico-pathologic features and peripheral blood inflammatory indexes were developed.

Result.

Molecular Subtype, HALP, P53, and FAR were used to construct the predictive models for traditional non-pCR (T-NpCR) and total-pCR (TpCR). The T-NpCR group was divided into DpCR and non-pCR (NpCR) subgroups to construct a new model to more accurately predict NAT outcomes. cN, HALP, FAR, Molecular Subtype, and RMC were used to construct the predictive models for NpCR and DpCR. The receiver operating characteristic (ROC) curves indicate that the model exhibits robust predictive capacity. Clinical Impact Curves (CIC) and Decision Curve Analysis (DCA) indicate that the models present a superior clinical utility.

Conclusion.

HALP and FAR were identified as peripheral blood inflammatory index predictors for accurately predicting NAT outcomes.

Breast Cancer

Neoadjuvant Therapy

Pathological Response

FAR

HALP

Nomogram

Lymph node metastasis constitutes a pivotal determinant in the recurrence and metastatic progression of breast cancer. Neoadjuvant therapy (NAT) has become the preferred treatment for node-positive breast cancer^[1]. Total-pathological complete response (TpCR) is defined as the absence of significant residual tumor in both the breast and lymph nodes by post-NAT pathological examination, serving as a critical indicator of favorable prognosis^[2]. However, previous studies have indicated the potential for discrepancies in the response of breast lesions and lymph nodes to neoadjuvant therapy. The American College of Surgeons Oncology Group (ACOSOG) reported a pCR rate of 27.8% in the breast and axilla in the Z1071 study, compared with only 5.7% in the breast alone and 13.1% in the lymph nodes alone^[3]. This phenomenon, where the tumor disappears in only one part of the breast and axillary lymph nodes, while remaining in another part, is known as a discordant pathological complete response (DpCR). Previous studies have generally considered DpCR as not having achieved complete pathological remission and therefore classified them into the NpCR group. However, in fact, patients with DpCR represent a unique prognostic group whose prognosis is intermediate between that of TpCR and non-pathological complete remission (NpCR) group^{[4, 5]}. DpCR is a transitional state which could partly response to neoadjuvant therapy. It suggests that it may be possible to convert DpCR to TpCR^[6].Therefore, DpCR patients should be studied in isolation from traditional NpCR (T-NpCR).^[7–9] A more precise estimation of the pathological response of breast cancer prior to NAT has significant implications for the assessment of disease risk and the development of therapeutic strategies. Research on the clinic features and prediction model for DpCR status is currently insufficient, requiring further scientific investigations to address this deficiency.

Breast cancer cells can interact with peripheral stromal and inflammatory cells to form inflammatory tumor microenvironment, which contributes to tumorigenesis, progression, invasion and chemoresistance, affecting patient prognosis^[10]. Peripheral blood inflammatory index including platelet-to-lymphocyte ratio (PLR), neutrophil-to-lymphocyte ratio (NLR), lymphocyte-to-monocyte ratio (LMR), and systemic immuno-inflammatory index (SII) exhibit a significant correlation with the prognosis of BC patients^[11–15]. Numerous retrospective clinical trials in breast cancer have used peripheral blood inflammatory indices to predict TpCR in NAT^[16]. However, there is a lack of research investigating the predictive potential of peripheral blood inflammatory index for DpCR.

This study aimed to identify clinic features and peripheral blood inflammatory indexes that could be used to predict pathological response degree following NAT. We developed two prediction models and validated the predictive ability of the models.

Patient Selection

Female patients with breast cancer who were treated with NAT from January 2011 to December 2023 at Harbin Medical University Cancer Hospital were recruited into this retrospective study. The inclusion criteria are as follows: (1) all patients underwent ultrasound-guided aspiration biopsy of the breast and associated lymph nodes before NAT. The biopsy pathology confirmed the presence of tumour metastases in these lymph nodes; (2) received standard neoadjuvant treatment program. All neoadjuvant regimens are based on NCCN guidelines; (3) Receive surgical resection and axillary lymph node dissection or sentinel lymph node biopsy. Patients with either of the following situations were excluded: (1) history of prior or concurrent breast cancer; (2) evidence of distant metastasis; (3) receiving less than two cycles of NAT; (4) abandoning NAT in the middle of process due to intolerable side-effects.

This study was approved by the Clinical Research Ethics Committee of the Harbin Medical University Cancer Hospital. This research complies with the 1964 World Medical Association Declaration of Helsinki and subsequently amended versions. All patients were provided with written informed consent prior to participation.

Definition of different pathological reaction statuses after NAT

All samples were evaluated using the AJCC ypTNM criteria. TpCR was defined as postoperative pathology in which no evidence of tumor residue was found in either the breast lesion or the lymph node (ypT0/ypN0). DpCR was defined as postoperative pathology in which tumor residue was found in one of the two, but not the other, in either the breast lesion or the lymph node (ypT ≥ 1/ypN0 or ypT0/ypN ≥ 1). Breast pCR or lymph node pCR are collectively known as DpCR, as previous studies have shown that they have the same survival^[7]. NpCR was defined as postoperative pathology in which tumor residue was found in both the breast lesion and the lymph node. The combination of DpCR and NpCR was described as T-NpCR, which is the traditional meaning of not achieving pCR after NAT. In addition, previous studies have demonstrated that survival rates for ductal carcinoma in situ (DCIS) or minimal residual disease (MRD) in the breast or lymph nodes show no significant difference from those observed for TpCR. Consequently, these two conditions have been included in the TpCR group^{[6, 17]}.

Pathologic features

Referring to the 2020 edition of the ASCO guidelines, an IHC score of ≥ 1% for ER or PR was defined as positive. IHC score of 0 or 1 + for HER2 was considered negative. In cases with an IHC score of 2+, fluorescence in situ hybridization (FISH) must be performed in addition to IHC. IHC score of 3 (> 10% of cells showing high-intensity periplasmic staining) or FISH positivity was considered positive. Ki-67 was defined as low expression if the proportion of positively stained cells was less than 15%. It was defined as moderate expression when the percentage of positively stained cells was between 15–29%. The high expression was defined when the percentage of positively stained cells was ≥ 29%.

Peripheral blood inflammation index

All peripheral blood inflammation index were calculated based on the haematology of the patients 7 days before they underwent NAT. The calculation of each index of peripheral blood inflammation was as follow: NLR (Neutrophil count/ Lymphocyte count), Pan-Immune-Inflammation-value (PIV, Platelet count ×Neutrophil count × Monocytes count/ Lymphocyte count), Systemic inflammatory response index (SIRI, Neutrophil count ×Monocyte count/ Lymphocyte count), the Hemoglobin, Albumin, Lymphocyte, Platelet Score (HALP, Hemoglobin count × Albumin × Lymphocyte count/ Platelet count ), the Fibrinogen-Albumin Ratio (FAR, Fibrinogen×100/ Albumin), the Fibrinogen-Platelet Ratio (FPR, Fibrinogen/ Platelet count), the Systemic Immunoinflammatory Index (SII, Platelet count × Neutrophil count/ Lymphocyte count), LMR (Lymphocyte count/ Monocyte count), PLR (Platelet count/ Lymphocyte count).

Follow-up

All patients received regular follow-up after completing treatment according to clinical guidelines. The follow-up programme was conducted at the following intervals: every three to four months for the initial two years, every six months for the subsequent three years, and annually thereafter. Participants were defined as having lost to follow-up if there was an interval of more than two years between the last visit and the final follow-up. All patients followed up until 30 January 2024. Disease-free survival (DFS) was defined as the interval between diagnosis and either breast cancer recurrence, death from any cause, or the date of the final follow-up. Similarly, overall survival (OS) was defined as the time from diagnosis to either death or the date of the final follow-up.

Statistical analysis

This study applied RStudio (version 4.3.2; https://www.r-project.org/) software for statistical analysis. The correlation between different groups was analyzed using the chi-square test. Lasso regression was employed to identify predictive features based on non-zero coefficients. The optimal parameter (lambda) selection in the Lasso model was cross-validated ten-fold based on the minimum criterion. The receiver operating characteristic curve (ROC) was employed to determine the best cut-off value. Kaplan-Meier curves and log-rank tests were employed to plot and compare the DFS and OS curves. The patients were randomly divided into a training and validation set in a 7:3 ratio. Lasso regression analysis, uni/multivariate logistic regression analysis were performed to identify factors associated with pCR status after NAT. P < 0.05 was considered a statistically significant.

Baseline of patients achieve T-NpCR and TpCR

A total of 789 patients were included in the study. Of 207 (26.24%) achieved TPCR, while 582 (73.76%) were evaluated for T-NpCR. The baseline clinic features and peripheral blood inflammatory indexes of patients are presented in Table 1. A chi-square test was performed for all variables. Molecular Subtype, Ki67, P53, NAT cycle, NAT regimen, HER2 Targeted Therapy, Carboplatin, SII, NLR, PLR, PIV, REC, RNC, RMC, FPR, FAR, HALP was found to be significantly different between the two groups. Therefore, these 17 variables were included in the screening process of predictors.

Table 1

Clinical baseline characteristics between T-NpCR and TpCR patient.
Characteristic	Levels	T-NpCR (N = 582)	TpCR (N = 207)	P
Age	> 60	104 (17.9%)	29 (14%)	.244
	≤ 60	478 (82.1%)	178 (86%)
BMI	Obese	40 (6.9%)	11 (5.3%)	.735
	Overweight	198 (34%)	72 (34.8%)
	Lean	344 (59.1%)	124 (59.9%)
Menstrual state	Menopause	320 (55%)	106 (51.2%)	.393
	Premenopause	262 (45%)	101 (48.8%)
Lesion number	Mono	462 (79.4%)	171 (82.6%)	.368
	Multi	120 (20.6%)	36 (17.4%)
cT	1 + 2	476 (81.8%)	179 (86.5%)	.151
	3 + 4	106 (18.2%)	28 (13.5%)
cN	1 + 2	381 (65.5%)	126 (60.9%)	.271
	3	201 (34.5%)	81 (39.1%)
Molecular Subtype	HR(+)HER2(-)	343 (58.9%)	26 (12.6%)	< .001
	HR(+)HER2(+)	74 (12.7%)	37 (17.9%)
	HR(-)HER2(+)	66 (11.3%)	68 (32.9%)
	HR(-)HER2(-)	99 (17%)	76 (36.7%)
Ki67	≥ 30%	287 (49.3%)	147 (71%)	< .001
	15% − 30%	172 (29.6%)	41 (19.8%)
	< 15%	123 (21.1%)	19 (9.2%)
P53	0	247 (42.4%)	141 (68.1%)	< .001
	1	219 (37.6%)	21 (10.1%)
	2	53 (9.1%)	15 (7.2%)
	3	63 (10.8%)	30 (14.5%)
NAT cycle	> 6	182 (31.3%)	74 (35.7%)	.273
	≤ 6	400 (68.7%)	133 (64.3%)
HER2 Targeted Therapy	Yes	110 (18.9%)	72 (34.8%)	< .001
	No	472 (81.1%)	135 (65.2%)
PDW	> 13.45	257 (44.2%)	84 (40.6%)	.417
	≤ 13.45	325 (55.8%)	123 (59.4%)
SIRI	> 1.08	130 (22.3%)	54 (26.1%)	.317
	≤ 1.08	452 (77.7%)	153 (73.9%)
SII	> 839.22	112 (19.2%)	20 (9.7%)	.002
	≤ 839.22	470 (80.8%)	187 (90.3%)
NLR	> 1.51	437 (75.1%)	139 (67.1%)	.034
	≤ 1.51	145 (24.9%)	68 (32.9%)
PLR	> 162.59	201 (34.5%)	49 (23.7%)	.005
	≤ 162.59	381 (65.5%)	158 (76.3%)
LMR	> 3.12	521 (89.5%)	191 (92.3%)	.313
	≤ 3.12	61 (10.5%)	16 (7.7%)
PIV	> 462.21	63 (10.8%)	12 (5.8%)	.048
	≤ 462.21	519 (89.2%)	195 (94.2%)
REC	> 0.87	349 (60%)	145 (70%)	.013
	≤ 0.87	233 (40%)	62 (30%)
RLC	> 37	134 (23%)	60 (29%)	.106
	≤ 37	448 (77%)	147 (71%)
RNC	> 57.69	373 (64.1%)	111 (53.6%)	.010
	≤ 57.69	209 (35.9%)	96 (46.4%)
RMC	> 6.32	227 (39%)	102 (49.3%)	.013
	≤ 6.32	355 (61%)	105 (50.7%)
FPR	> 1.11	322 (55.3%)	76 (36.7%)	< .001
	≤ 1.11	260 (44.7%)	131 (63.3%)
FAR	> 6.6	363 (62.4%)	63 (30.4%)	< .001
	≤ 6.6	219 (37.6%)	144 (69.6%)
HALP	> 37.32	336 (57.7%)	160 (77.3%)	< .001
	≤ 37.32	246 (42.3%)	47 (22.7%)
T-NpCR: traditional non-pathologic complete response, DpCR: discordant-pathologic complete response, TpCR: total-pathologic complete response,BMI: body mass index, PDW: platelet distribution width, REC: relative eosinophilic count, RLC: relative lymphocyte count, RNC: relative neutrophil count, RMC: relative macrophage cell count.

Identification of predictors of the T-NpCR and TpCR group

All patients were randomly divided into validation set (n = 267) and training set (n = 522) according to a 7:3 ratio. Four parameters (molecular subtypes, P53, FAR, and HALP) were identified as potential predictors by Lasso (Fig. 1A-B) and uni/multivariate (Fig. 1C) logistic regression analysis. Based on these predictors, nomogram were constructed (Fig. 1D).

Evaluating the prediction performance of T-NpCR and TpCR group model in train set

The ROC curve showed that AUC value of the model was 0.803 (95% CI: 0.761–0.846), and the AUC values of Molecular Subtype, FAR, HALP and P53 were 0.728 (95%CI: 0.684–0.772), 0.701 (95%CI: 0.653–0.749), 0.612 (95%CI: 0.56–0.664), 0.602 (95%CI: 0.549–0.656) (Fig. 2A). The calibration curve demonstrated that the average absolute error of the model was 0.014, indicating that the model exhibited enhanced predictive capability (Fig. 2B). To assess the clinical utility and predictive capacity of nomograms, DCA and CIC were plotted. The models in the T-NpCR and TpCR group exhibited superior clinical utility when the threshold probability values were within the range of 0.35 to 0.77 (Fig. 2C-D).

Baseline and prognosis of patients achieve NpCR and DpCR

To further understand the prognosis of patients with different extent of pathological response after NAT for breast cancer, the T-NpCR cohort was divided into two subgroups: NpCR and DpCR. The mean follow-up period for the entire cohort was 49 months (3–88 months). Based on the follow-up data, K-M curves were plotted for patients in the three groups (Fig. 3A-B). The result demonstrated that patients in the TpCR group exhibited the best OS and DFS, followed by the DpCR group, and the NpCR group proved to be the worst. The observed differences in survival between each two groups were statistically significant. Thus, DpCR represents a subgroup with a distinct prognosis, making it scientifically meaningful to study it separately from T-NpCR. The baseline clinical characteristics of the NpCR and DpCR groups are shown in Table 2. Among them, 314 patients were assessed as NpCR, and 268 patients were assessed as DpCR. Chi-square tests revealed statistically significant differences between the groups for the variables Lesion number, cN, Molecular Subtype, P53, NAT regimen, HER2 targeted therapy, RMC, FPR, FAR, and HALP. These 10 variables were included in the predictor selection process for model construction. Five parameters (cN, Molecular Subtype, RMC, HALP and FAR) were identified as potential predictors by Lasso (Fig. 3C-D) and uni/multivariate (Fig. 3E) logistic regression analysis. Based on these predictors, nomogram were constructed (Fig. 3F).

Table 2

Clinical baseline characteristics between NpCR and DpCR group.
Characteristic	Levels	NpCR (N = 314)	DpCR (N = 268)	P
Age	> 60	58 (18.5%)	46 (17.2%)	.763
	≤ 60	256 (81.5%)	222 (82.8%)
BMI	Obese	20 (6.4%)	20 (7.5%)	.623
	Overweight	112 (35.7%)	86 (32.1%)
	Lean	182 (58%)	162 (60.4%)
Menstrual state	Menopause	182 (58%)	138 (51.5%)	.139
	Premenopause	132 (42%)	130 (48.5%)
Lesion number	Mono	234 (74.5%)	228 (85.1%)	.002
	Multi	80 (25.5%)	40 (14.9%)
cT	1	255 (81.2%)	221 (82.5%)	.778
	2	59 (18.8%)	47 (17.5%)
cN	1 + 2	190 (60.5%)	191 (71.3%)	.008
	3	124 (39.5%)	77 (28.7%)
Molecular Subtype	HR(+)HER2(-)	218 (69.4%)	125 (46.6%)	< .001
	HR(+)HER2(+)	27 (8.6%)	47 (17.5%)
	HR(-)HER2(+)	22 (7%)	44 (16.4%)
	HR(-)HER2(-)	47 (15%)	52 (19.4%)
Ki67	≥ 30%	144 (45.9%)	143 (53.4%)	.192
	15% − 30%	100 (31.8%)	72 (26.9%)
	< 15%	70 (22.3%)	53 (19.8%)
P53	0	133 (42.4%)	114 (42.5%)	.003
	1	127 (40.4%)	84 (31.3%)
	2	20 (6.4%)	41 (15.3%)
	3	34 (10.8%)	29 (10.8%)
NAT cycle	> 8	70 (22.3%)	86 (32.1%)	.010
	≤ 8	244 (77.7%)	182 (67.9%)
HER2 targeted therapy	Yes	51 (16.2%)	64 (23.9%)	.028
	No	263 (83.8%)	204 (76.1%)
PDW	> 13.05	138 (43.9%)	119 (44.4%)	.979
	≤ 13.05	176 (56.1%)	149 (55.6%)
SIRI	> 0.54	74 (23.6%)	56 (20.9%)	.502
	≤ 0.54	240 (76.4%)	212 (79.1%)
SII	> 406.8	62 (19.7%)	50 (18.7%)	.821
	≤ 406.8	252 (80.3%)	218 (81.3%)
NLR	> 1.96	236 (75.2%)	201 (75%)	1.000
	≤ 1.96	78 (24.8%)	67 (25%)
PLR	> 137.55	107 (34.1%)	94 (35.1%)	.869
	≤ 137.55	207 (65.9%)	174 (64.9%)
LMR	> 6.01	281 (89.5%)	240 (89.6%)	1.000
	< 6.01	33 (10.5%)	28 (10.4%)
PIV	> 130.77	34 (10.8%)	29 (10.8%)	1.000
	≤ 130.77	280 (89.2%)	239 (89.2%)
REC	> 1.14	185 (58.9%)	164 (61.2%)	.636
	≤ 1.14	129 (41.1%)	104 (38.8%)
RLC	> 44.82	75 (23.9%)	59 (22%)	.663
	≤ 44.82	239 (76.1%)	209 (78%)
RNC	> 61.26	204 (65%)	169 (63.1%)	.695
	≤ 61.26	110 (35%)	99 (36.9%)
RMC	> 6.33	110 (35%)	117 (43.7%)	.041
	≤ 6.33	204 (65%)	151 (56.3%)
FPR	> 0.95	198 (63.1%)	124 (46.3%)	< .001
	≤ 0.95	116 (36.9%)	144 (53.7%)
FAR	> 6.57	242 (77.1%)	121 (45.1%)	< .001
	≤ 6.57	72 (22.9%)	147 (54.9%)
HALP	> 53.15	195 (62.1%)	141 (52.6%)	.026
	≤ 53.15	119 (37.9%)	127 (47.4%)
NpCR: non-pathologic complete response, DpCR: discordant-pathologic complete response.

Evaluating the prediction performance of NpCR and DpCR subgroups model in train set

The ROC curve demonstrated that the overall AUC value of the model was 0.74 (95% CI: 0.691–0.788). The AUC values for the variables cN, HALP, FAR, Molecular Subtype, and RMC were 0.559 (95% CI: 0.5136–0.6042), 0.6232 (95% CI: 0.5748–0.6715), 0.5663 (95% CI: 0.5106–0.6221), 0.6023 (95% CI: 0.5481–0.6565), and 0.6665 (95% CI: 0.6132–0.7198), respectively (Fig. 4A). The calibration curve showed that the average absolute error of the model was 0.044, indicating a relatively reliable predictive capability (Fig. 4B). The DCA and CIC analysis demonstrated that the model exhibited good clinical utility for the NpCR and DpCR subgroups when the threshold probability was larger than 0.43 (Fig. 4C-D).

Evaluation of models prediction performance in validation set

The formulas for the two multivariate logistic regression models were derived and their predictive performance was evaluated in the validation set. The ROC curve demonstrated an AUC of 0.7437 (95% CI: 0.6638–0.8236) for the T-NpCR vs TpCR group and 0.798 (95% CI: 0.7273–0.8704) for the NpCR vs DPCR group (Fig. 5A, E). Calibration curves based on the validation set data showed that the observed value curves were close to the actual values (Fig. 5B, F). Additionally, DCA and CIC curves indicated that nomograms of the models of T-NpCR vs TpCR group (Fig. 5C, D) and NpCR vs DpCR group (Fig. 5J, H) provided high predictive accuracy for the pathological response status of patients treated with NAT.

This study included 789 patients with breast-related lymph node metastases at baseline status who underwent NAT. Molecular Subtype, HALP, P53, and FAR were identified as predictors of T-NpCR and TpCR groups by regression analysis, and prediction models were constructed and validated. The T-NpCR group was divided into two subgroups, NpCR and DpCR, for subgroup analyses. Kaplan-Meier curves demonstrated that the TpCR group had the greatest survival for OS and DFS, followed by the DpCR group, while the NpCR group had the worst. NpCR and DpCR prediction models based on cN, HALP, FAR, molecular subtype and RMC were constructed and their predictive performance was verified in the training and validation sets.

Chen et al. showed that HR(-)HER2(+) subtype of breast cancer had the highest rate of BpCR and NpCR, and HR(+)HER2(-) subtype had the lowest rate of BpCR^[18]. This indicates that the heterogeneity of different subtypes of tumors, different biological behaviors, and NAT regimens contributing to DpCR. It has been suggested that cancer cells in lymph nodes may have a tumor immune tolerance, and potential explanations include differences in chemotherapy sensitivity of metastatic tumor cells or the protective effect of the lymph node microenvironment on the tumor^[19]. A study by Rene et al. showed that lymphatic dysfunction were more likely to have DpCR^[20]. Previous studies have indicated that fibrinogen deposition, diminished immune response, or combination of chronic systemic diseases (e.g., diabetes mellitus) may contribute to lymphatic dysfunction. This ultimately results in the inadequate delivery of NAT drugs within the lymphatic system, or in the failure of to interact with tumor foci^{[21, 22]}. In this study, elevating peripheral blood fibrinogen was associated with a worse pathological response status. The two predictive models developed in this study indicated that molecular subtype of breast cancer was associated with different pathological response status. The analysis of the baseline characteristics revealed that patients with HER2-positive or triple-negative types were more likely to have better pathological response. Nevertheless, the precise mechanism remains to be elucidated through further investigation. To ascertain the mechanisms underlying DpCR, more in-depth studies are needed to identify which neoadjuvant treatment strategies may transform DpCR into TpCR status. Additionally, the axillary lymph node management strategies employed for patients undergoing NAT prove pivotal in enhancing long-term survival and reducing recurrence in breast cancer patients.^[23]

Peripheral blood immune cells can partially respond to the inflammatory state in the immune microenvironment, which is the theoretical basis for the hypothesis that peripheral blood inflammatory markers may predict tumor prognosis^[24]. Abnormalities in coagulation can increase the risk of thrombosis and have a pro-tumorigenic effect, and indicators such as albumin and haemoglobin can reflect the overall nutritional status of the patient^[25–27]. Peripheral blood inflammatory indexes have been shown in several studies to potentially predict prognostic status or NAT outcome in breast cancer^{[28, 29]}. However, the capability of these peripheral blood parameters to predict between the three different pCR status of NpCR, DpCR, and TpCR is still unclear. Our study screened peripheral blood inflammatory indexes and clinico-pathological features that might predict pathological response status after NAT by lasso regression, univariate and multivariate regression analyses, and screened parameters that had good ability to predict the three pathological response statuses two by two. Based on these parameters we plotted nomograms, and the AUC values of the two model groups were as follows: in the T-NpCR vs TpCR group: 0.803 (95% CI: 0.761–0.846) and in the NpCR vs DpCR group: 0.74 (95% CI: 0.691–0.788). In addition, we noticed that HALP and FAR appeared in both predictive models and had a longer share of the scoring axis in the nomogram compared to other predictors in the same group. This suggests that the combination of HALP and FAR may have a better ability to predict the extent of tumor remission after NAT in breast cancer.

The HALP integrates four routinely collected indicators of immune and nutritional status and has been used as a new prognostic biomarker to predict many clinical outcomes in a variety of tumors. A meta analysis that included tumors such as gastric and cervical cancers showed that low HALP at baseline status was associated with poor prognosis of the tumor^[30]. Lou et al. demonstrated that baseline HALP could be a predictor of whether or not to pCR after NAT in breast cancer. Using a cut-off value of 24.14, the OR for low HALP was 0.518 (95% CI: 0.365–0.734), and the area under the ROC curve for HALP was 0.847^[31]. Another 2022 study discussed whether HALP could be used as a predictor for the presence or absence of axillary lymph node involvement, and demonstrated that the rate of axillary lymph node involvement for HALP less than 29.01 was 67.7% and 53.3% for HALP greater than or equal to 29.01 (p = 0.038)^[32]. In our study, the level of HALP was significantly lower in the NpCR group than in the TpCR group (45.7 ± 19.2 VS 56.6 ± 24.2). This phenomenon is consistent with previous studies. Interestingly, the HALP level in the DpCR group were lower than which in the NpCR group (45.7 ± 19.2 VS 40.9 ± 14.7) and the difference was statistically significant. The results of univariate and multivariate regression analyses also matched this trend. This suggests that the relationship between HALP score and prognosis may not be strictly positive. Also, we noted that platelets were highest in the DpCR group, followed by the NpCR group, and smallest in the TpCR group. The peripheral blood inflammation indexes associated with platelets were broadly consistent with this trend. Apart from platelets, haemoglobin, albumin and lymphocytes could not explain this trend. This implies that platelets and the coagulation system may play a role in the formation of DpCR, making neoadjuvant therapy less responsive in a subset of patients who may achieve TpCR or making oncological treatment slightly more effective in patients who may NpCR.

FAR is a coagulation-inflammation-nutritional indicator of prognosis in a variety of solid tumors^[33–37]. Since infection, blood coagulation, and so on affect plasma fibrinogen values, fibrinogen can somewhat represent the degree of inflammatory response^[38]. Hwang et al. showed that patients with high FAR (cut-off value of 7.1) had a worse prognosis, and that univariate ( HR: 2.722, 95% CI: 1.659–4.468, P < 0.001) and multivariate (HR: 2.622, 95% CI: 1.455–4.724, P = 0.001) regression analyses also confirmed this^[39]. Yang et al. set the cut-off value of FAR at 6.6 in their study, and survival analyses showed that high FAR implied worse OS and DFS^[40]. In contrast, however, Zheng et al. reached the opposite result. This study concluded that low FAR (≤ 8.4) was protective for patients and that OS and DFS were worse with high FAR-PLR scores^[41]. To date there are no studies discussing whether FAR can predict DpCR status after NAT. In our study, it was found that low FAR predict patients with better NAT responsiveness. The cut-off values of FAR were calculated from ROC curves to be 6.572 (NpCR vs. DpCR) and 5.513 (NpCR vs. TpCR) in the two groups, respectively. Unlike HALP, there was a more direct correlation between FAR and NAT outcome. That is, lower FAR indicate a better pathological response. Based on our findings, the coagulation system-related components, especially platelets and fibrinogen, may correlate with pathological response status after NAT. However, the mechanisms behind these findings still need to be supplemented and validated by further research.

Despite the encouraging results, this study has several limitations: (1) There is a lack of an external validation cohort to test the conclusions; (2) The findings of this retrospective study should be validated by further prospective studies. What’s more, further validation in larger cohorts is required before the models can be applied in routine clinical practice.

In conclusion, this study identified potential factors affecting the outcome of NAT in breast cancer. FAR and HALP were found to be potential indicators that could be used to accurately predict pathological responses to NAT in breast cancer.

Pathological complete response

pCR

Neoadjuvant therapy

NAT

Traditional NpCR

T-NpCR

Total

pathological complete response-TpCR

Discordant pathological complete response

DpCR

Non pCR

NpCR

American College of Surgeons Oncology Group

ACOSOG

Platelet

to-lymphocyte ratio-PLR

Neutrophil

to-lymphocyte ratio-NLR

Lymphocyte

to-monocyte ratio-LMR

Systemic immuno

inflammatory index-SII

Ductal carcinoma in situ

DCIS

Minimal residual disease

MRD

Fluorescence in situ hybridization

FISH

Pan

Immune-Inflammation-value-PIV

Systemic inflammatory response index

SIRI

The Hemoglobin, Albumin, Lymphocyte, Platelet Score

HALP

Fibrinogen

Albumin Ratio-FAR

Fibrinogen

Platelet Ratio-FPR

Disease

free survival-DFS

Overall survival

Receiver operating characteristic curve

ROC

Clinical Impact Curves

CIC

Decision Curve Analysis

DCA

Acknowledgements

Not applicable.

Authors’ contributions

Shuqiang Liu, Cong Jiang, Shiyuan Zhang, Kun Qiao and Yuanxi Huang contributed to the concept and design of the study. Danping Wu, Xiaotian Yang and Boqian Yu contributed to the acquisition and interpretation of data and drafting the article. All authors read and approved the final version of the article.

Funding

The authors did not receive support from any organization for the submitted work.

Availability of data and materials

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Ethical approval and consent to participate

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Author details

¹Department of Breast Surgery, Harbin Medical University Cancer Hospital, Harbin, China, ²Department of Breast Surgery, The Third Affiliated Hospital of Kunming Medical University, Kunming, China

Conforti F, Pala L, Sala I, et al. Evaluation of pathological complete response as surrogate endpoint in neoadjuvant randomised clinical trials of early stage breast cancer: systematic review and meta-analysis. BMJ. 2021;375:e066381.
Long-term outcomes for. neoadjuvant versus adjuvant chemotherapy in early breast cancer: meta-analysis of individual patient data from ten randomised trials. Lancet Oncol. 2018;19(1):27–39.
Boughey JC, Ballman KV, McCall LM, et al. Tumor Biology and Response to Chemotherapy Impact Breast Cancer-specific Survival in Node-positive Breast Cancer Patients Treated With Neoadjuvant Chemotherapy: Long-term Follow-up From ACOSOG Z1071 (Alliance). Ann Surg. 2017;266(4):667–76.
Hennessy BT, Hortobagyi GN, Rouzier R, et al. Outcome after pathologic complete eradication of cytologically proven breast cancer axillary node metastases following primary chemotherapy. J Clin Oncol. 2005;23(36):9304–11.
Bhargava R, Beriwal S, Dabbs DJ, et al. Immunohistochemical surrogate markers of breast cancer molecular classes predicts response to neoadjuvant chemotherapy: a single institutional experience with 359 cases. Cancer. 2010;116(6):1431–9.
Go J, Ahn JH, Park JM, et al. Distinct Prognosis of Minimal Residual Disease According to Breast Cancer Subtype in Patients with Breast or Nodal Pathologic Complete Response After Neoadjuvant Chemotherapy. Ann Surg Oncol. 2023;30(12):7060–8.
Fayanju OM, Ren Y, Thomas SM, et al. The Clinical Significance of Breast-only and Node-only Pathologic Complete Response (pCR) After Neoadjuvant Chemotherapy (NACT): A Review of 20,000 Breast Cancer Patients in the National Cancer Data Base (NCDB). Ann Surg. 2018;268(4):591–601.
Lee MK, Srour MK, Walcott-Sapp S, et al. Impact of the extent of pathologic complete response on outcomes after neoadjuvant chemotherapy. J Surg Oncol. 2020;121(2):210–5.
Samiei S, Simons JM, Engelen S, Beets-Tan R, Classe JM, Smidt ML. Axillary Pathologic Complete Response After Neoadjuvant Systemic Therapy by Breast Cancer Subtype in Patients With Initially Clinically Node-Positive Disease: A Systematic Review and Meta-analysis. JAMA Surg. 2021;156(6):e210891.
Tanaka A, Sakaguchi S. Regulatory T cells in cancer immunotherapy. Cell Res. 2017;27(1):109–18.
Qi Q, Zhuang L, Shen Y, et al. A novel systemic inflammation response index (SIRI) for predicting the survival of patients with pancreatic cancer after chemotherapy. Cancer. 2016;122(14):2158–67.
Chen L, Kong X, Yan C, Fang Y, Wang J. The Research Progress on the Prognostic Value of the Common Hematological Parameters in Peripheral Venous Blood in Breast Cancer. Onco Targets Ther. 2020;13:1397–412.
Fucà G, Guarini V, Antoniotti C, et al. The Pan-Immune-Inflammation Value is a new prognostic biomarker in metastatic colorectal cancer: results from a pooled-analysis of the Valentino and TRIBE first-line trials. Br J Cancer. 2020;123(3):403–9.
Rimini M, Casadei-Gardini A, Ravaioli A, et al. Could Inflammatory Indices and Metabolic Syndrome Predict the Risk of Cancer Development? Analysis from the Bagnacavallo Population Study. J Clin Med. 2020;9(4):1177.
Lee SY, Lee SI, Min BW, Oh SC. Prognostic implication of systemic inflammatory markers in young patients with resectable colorectal cancer. Ann Surg Treat Res. 2021;100(1):25–32.
Xie J, Guo Z, Zhu Y, Ma M, Jia G. Peripheral blood inflammatory indexes in breast cancer: A review. Med (Baltim). 2023;102(48):e36315.
Jones RL, Lakhani SR, Ring AE, Ashley S, Walsh G, Smith IE. Pathological complete response and residual DCIS following neoadjuvant chemotherapy for breast carcinoma. Br J Cancer. 2006;94(3):358–62.
Chen SC, Yu CC, Chang HK, Lin YC, Huang YT. Discrepancy of Breast and Axillary Pathologic Complete Response and Outcomes in Different Subtypes of Node-positive Breast Cancer after Neoadjuvant Chemotherapy. J Cancer. 12(17): 5365–74.
Reticker-Flynn NE, Zhang W, Belk JA, et al. Lymph node colonization induces tumor-immune tolerance to promote distant metastasis. Cell. 2022;185(11):1924–e194223.
Flores R, Roldan E, Pardo JA, et al. Discordant Breast and Axillary Pathologic Response to Neoadjuvant Chemotherapy. Ann Surg Oncol. 2023;30(13):8302–7.
Costantini V, Zacharski LR, Memoli VA, et al. Fibrinogen deposition and macrophage-associated fibrin formation in malignant and nonmalignant lymphoid tissue. J Lab Clin Med. 1992;119(2):124–31.
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(2):460–9.
Bundred NJ, Barnes NL, Rutgers E, Donker M. Is axillary lymph node clearance required in node-positive breast cancer. Nat Rev Clin Oncol. 2015;12(1):55–61.
Jalali A, Miresse D, Fahey MR, et al. Peripheral Blood Cell Ratios as Prognostic Indicators in a Neoadjuvant Chemotherapy-Treated Breast Cancer Cohort. Curr Oncol. 2022;29(10):7512–23.
Neman J, Termini J, Wilczynski S, et al. Human breast cancer metastases to the brain display GABAergic properties in the neural niche. Proc Natl Acad Sci U S A. 2014;111(3):984–9.
So CL, Saunus JM, Roberts-Thomson SJ, Monteith GR. Calcium signalling and breast cancer. Semin Cell Dev Biol. 2019;94:74–83.
Zhang C, Yang Z, Zhou P, et al. Phosphatidylserine-exposing tumor-derived microparticles exacerbate coagulation and cancer cell transendothelial migration in triple-negative breast cancer. Theranostics. 2021;11(13):6445–60.
Yuan Y, Vora N, Sun CL, et al. Association of pre-chemotherapy peripheral blood pro-inflammatory and coagulation factors with reduced relative dose intensity in women with breast cancer. Breast Cancer Res. 2017;19(1):101.
Yamanouchi K, Maeda S. The Efficacy of Inflammatory and Immune Markers for Predicting the Prognosis of Patients with Stage IV Breast Cancer. Acta Med Okayama. 2023;77(1):37–43.
Xu H, Zheng X, Ai J, Yang L. Hemoglobin, albumin, lymphocyte, and platelet (HALP) score and cancer prognosis: A systematic review and meta-analysis of 13,110 patients. Int Immunopharmacol. 2023;114:109496.
Lou C, Jin F, Zhao Q, Qi H. Correlation of serum NLR, PLR and HALP with efficacy of neoadjuvant chemotherapy and prognosis of triple-negative breast cancer. Am J Transl Res. 2022;14(5):3240–6.
Duran A, Pulat H, Cay F, Topal U. Importance of HALP Score in Breast Cancer and its Diagnostic Value in Predicting Axillary Lymph Node Status. J Coll Physicians Surg Pak. 2022;32(6):734–9.
Acharya P, Jakobleff WA, Forest SJ, et al. Fibrinogen Albumin Ratio and Ischemic Stroke During Venoarterial Extracorporeal Membrane Oxygenation. ASAIO J. 2020;66(3):277–82.
Deng S, Fan Z, Xia H, et al. Fibrinogen/Albumin Ratio as a Promising Marker for Predicting Survival in Pancreatic Neuroendocrine Neoplasms. Cancer Manag Res. 2021;13:107–15.
Chen W, Shan B, Zhou S, Yang H, Ye S. Fibrinogen/albumin ratio as a promising predictor of platinum response and survival in ovarian clear cell carcinoma. BMC Cancer. 2022;22(1):92.
Wang S, Feng Y, Xie Y, et al. High fibrinogen-albumin ratio index (FARI) predicts poor survival in head and neck squamous cell carcinoma patients treated with surgical resection. Eur Arch Otorhinolaryngol. 2022;279(9):4541–8.
Li Q, Zhang J, Gao Q, et al. Preoperative Fibrinogen Albumin Ratio is an Effective Biomarker for Prognostic Evaluation of Gallbladder Carcinoma After Radical Resection: A 10-Year Retrospective Study at a Single Center. J Inflamm Res. 2023;16:677–89.
Zhang J, Ding Y, Wang W et al. Combining the Fibrinogen/Albumin Ratio and Systemic Inflammation Response Index Predicts Survival in Resectable Gastric Cancer. Gastroenterol Res Pract. 2020. 2020: 3207345.
Hwang KT, Chung JK, Roh EY, et al. Prognostic Influence of Preoperative Fibrinogen to Albumin Ratio for Breast Cancer. J Breast Cancer. 2017;20(3):254–63.
Yang Q, Liang D, Yu Y, Lv F. The Prognostic Significance of the Fibrinogen-to-Albumin Ratio in Patients With Triple-Negative Breast Cancer: A Retrospective Study. Front Surg. 2022;9:916298.
Zheng Y, Wu C, Yan H, Chen S. Prognostic value of combined preoperative fibrinogen-albumin ratio and platelet-lymphocyte ratio score in patients with breast cancer: A prognostic nomogram study. Clin Chim Acta. 2020;506:110–21.

No competing interests reported.

Download PDF

Journal Publication

published 12 Oct, 2024

Read the published version in BMC Women's Health →

Editorial decision: Revision requested
23 Aug, 2024
Reviews received at journal
21 Aug, 2024
Reviewers agreed at journal
21 Aug, 2024
Reviews received at journal
21 Aug, 2024
Reviewers agreed at journal
21 Aug, 2024
Reviewers agreed at journal
21 Aug, 2024
Reviewers agreed at journal
21 Aug, 2024
Reviewers agreed at journal
20 Aug, 2024
Reviewers agreed at journal
20 Aug, 2024
Reviewers agreed at journal
20 Aug, 2024
Reviewers agreed at journal
20 Aug, 2024
Reviews received at journal
20 Aug, 2024
Reviewers agreed at journal
20 Aug, 2024
Reviewers agreed at journal
20 Aug, 2024
Reviewers invited by journal
20 Aug, 2024
Editor invited by journal
20 Aug, 2024
Editor assigned by journal
15 Aug, 2024
Submission checks completed at journal
15 Aug, 2024
First submitted to journal
15 Aug, 2024

You are reading this latest preprint version

Development of Predictive Models for Pathological Response Status in Breast Cancer after Neoadjuvant Therapy Based on Peripheral Blood Inflammatory Indexes

Status:

Journal Publication

Version 1

Abstract

Background.

Method.

Result.

Conclusion.

Figures

Introduction

Methods and materials

Patient Selection

Definition of different pathological reaction statuses after NAT

Pathologic features

Peripheral blood inflammation index

Follow-up

Statistical analysis

Results

Baseline of patients achieve T-NpCR and TpCR

Identification of predictors of the T-NpCR and TpCR group

Evaluating the prediction performance of T-NpCR and TpCR group model in train set

Baseline and prognosis of patients achieve NpCR and DpCR

Evaluating the prediction performance of NpCR and DpCR subgroups model in train set

Evaluation of models prediction performance in validation set

Discussion

Conclusion

Abbreviations

Declarations

References

Additional Declarations

Status:

Journal Publication

Version 1