Improving prognosis of surrogate assay for breast cancer patients by absolute quantitation of Ki67 protein levels using Quantitative Dot Blot (QDB) method

Abstract

Purpose: The separation of Luminal A-like from Luminal B-like breast cancer subtypes in surrogate assay relies on Ki67 scores assessed by immunohistochemistry (IHC), a method known to be associated with subjectivity and inconsistency. We attempted to measure Ki67 levels absolutely, quantitatively and objectively in Formalin Fixed Paraffin Embedded (FFPE) specimens, and evaluate its influence on the performance of surrogate assay for breast cancer patients. Methods: The Ki67 protein levels were assessed using both IHC and Quantitative Dot Blot (QDB) methods respectively in 253 specimens. These patients were assigned into Luminal A-like and Luminal B-like subtypes using either Ki67 score of 14% as cutoff in surrogate assay, or 2.31 nmole/g from QDB method as cutoff in adjusted surrogate assay. These two subtyping methods were compared with the Kaplan-Meier, univariate and multivariate survival analyses of the overall survival (OS) of Luminal-like patients. Results: Ki67 levels measured using QDB method was highly correlated with those by IHC analysis (r=0.7, p<0.0001). The survival prediction for Luminal A-like patients was improved significantly in adjusted surrogate assay than surrogate assay (p=0.03 vs p<0.00052). The prediction of Hazard Ratio (HR) was also improve from 2.14 (95%CI: 0.89-5.11, p=0.087) to 6.89 (95%CI: 2.66-17.84, p<0.00001) in multivariate survival analysis. Conclusion: Our study demonstrated that the inherent subjectivity and inconsistency associated with IHC analysis has adverse effect on the performance of surrogate assay. This issue can be improved by objective and quantitative measurement of Ki67 levels with QDB method in daily clinical practice.

Combined use of absolutely quantitated cyclin D1 and Ki67 protein levels to improve prognosis of Luminal-like breast cancer

“…We found C l group have 10-year survival probability at 89%, in comparison to the 57% for C h group, with p<0.0001 from Log Rank test. As an comparison, the prognosis from adjusted surrogate assay, where the Luminal-like specimens were subtyped based on absolutely quantitated Ki67 protein levels, had the 10-year survival probability at 91% for Luminal A-like subtype, and 63% for Luminal B-like subtype, with p=0.00052[7]. For current prevailing surrogate assay, the 10-year survival probability was at 88% for Luminal A-like subtype, and 68% for Luminal B-like subtype with p=0.031 from Log Rank test using the same specimens.…”

“…These Luminal-like specimens were also separated into Ki67 high (K h ) and Ki67 low (K l ) group using the 2.31 nmole/g we defined in the previous study[7]. These two cutoffs were used together to further separate the Luminal-like specimens into four groups, with C h K h for specimens with the expression levels of both biomarkers above the respective cutoffs, C l K l for those below the respective cutoffs.…”

“…As shown in table 4, when both biomarkers were used, we identified C l K l group with the best prognosis (n=52), with 10-year survival probability at 94%, in comparison to the Luminal A-like subtype at 91% for adjusted surrogate assay (n=69), and 88% for surrogate assay (n=61); and C h K h group with the worst prognosis (n=34) with 10-year survival probability at 41%, in comparison to 63% for Luminal B-like subtype from adjusted surrogate assay (n=74) and 68% from surrogate assay (n=82) [7]. In addition, in C h K h group, the percentage of deceased patients were 47.05% (16/34), in comparison to 31.08% (23/74) in Luminal B-like subtype from adjusted surrogate assay and 26.83% (22/82) from surrogate assay.…”

“…We proposed to use cyclin D1 together with Ki67 protein biomarkers routinely in daily clinical practice to aid subtyping of breast cancer patients. ClKl: specimens with both cyclin D1 and Ki67 levels below the respective suggested cutoffs; ChKl: specimens with only Ki67 levels below the suggested cutoff; ClKh: specimens with only cyclin D1 below the suggested cutoff; ChKh: specimens with both cyclin D1 and Ki67 levels above the respective suggested cutoffs; Luminal Aq & Luminal Bq: Luminal A-like and B-like subtypes based on absolute quantitated Ki67 levels using QDB method, using 2.31 nmole/g as cutoff [7]; Luminal Ai & Bi: Luminal A-like and B-like subtypes based on Ki67 scores from IHC analysis, using 14% as cutoff [7].…”

“…In a previous study, we measured Ki67 protein levels absolutely and quantitatively in 143 Luminal-like Formalin Fixed Paraffin Embedded (FFPE) specimens using our independently developed Quantitative Dot Blot (QDB) method[7]. When replacing Ki67 score from IHC analysis with our absolutely quantitated Ki67 levels, we found the prognosis of surrogate assay was improved significantly for OS of Luminal-like tumors[7].…”