Limits...
Standardization for Ki-67 assessment in moderately differentiated breast cancer. A retrospective analysis of the SAKK 28/12 study.

Varga Z, Cassoly E, Li Q, Oehlschlegel C, Tapia C, Lehr HA, Klingbiel D, Thürlimann B, Ruhstaller T - PLoS ONE (2015)

Bottom Line: Fleiss'-kappa-values (14% cut-off) were the highest (moderate) using the original recommendation on light-microscope (Kappa 0.58).Fleiss' kappa values (20% cut-off) were the highest (Kappa 0.48 each) in analyzing hotspots on light-microscopy and digital-analysis.Region analysis and individual review (the original recommendation) on light-microscopy yielded the highest inter-observer reliability.

View Article: PubMed Central - PubMed

Affiliation: Institute of Surgical Pathology, University Hospital Zurich, Zurich, Switzerland.

ABSTRACT

Background: Proliferative activity (Ki-67 Labelling Index) in breast cancer increasingly serves as an additional tool in the decision for or against adjuvant chemotherapy in midrange hormone receptor positive breast cancer. Ki-67 Index has been previously shown to suffer from high inter-observer variability especially in midrange (G2) breast carcinomas. In this study we conducted a systematic approach using different Ki-67 assessments on large tissue sections in order to identify the method with the highest reliability and the lowest variability.

Materials and methods: Five breast pathologists retrospectively analyzed proliferative activity of 50 G2 invasive breast carcinomas using large tissue sections by assessing Ki-67 immunohistochemistry. Ki-67-assessments were done on light microscopy and on digital images following these methods: 1) assessing five regions, 2) assessing only darkly stained nuclei and 3) considering only condensed proliferative areas ('hotspots'). An individual review (the first described assessment from 2008) was also performed. The assessments on light microscopy were done by estimating. All measurements were performed three times. Inter-observer and intra-observer reliabilities were calculated using the approach proposed by Eliasziw et al. Clinical cutoffs (14% and 20%) were tested using Fleiss' Kappa.

Results: There was a good intra-observer reliability in 5 of 7 methods (ICC: 0.76-0.89). The two highest inter-observer reliability was fair to moderate (ICC: 0.71 and 0.74) in 2 methods (region-analysis and individual-review) on light microscopy. Fleiss'-kappa-values (14% cut-off) were the highest (moderate) using the original recommendation on light-microscope (Kappa 0.58). Fleiss' kappa values (20% cut-off) were the highest (Kappa 0.48 each) in analyzing hotspots on light-microscopy and digital-analysis. No methodologies using digital-analysis were superior to the methods on light microscope.

Conclusion: Our results show that all methods on light-microscopy for Ki-67 assessment in large tissue sections resulted in a good intra-observer reliability. Region analysis and individual review (the original recommendation) on light-microscopy yielded the highest inter-observer reliability. These results show slight improvement to previously published data on poor-reproducibility and thus might be a practical-pragmatic way for routine assessment of Ki-67 Index in G2 breast carcinomas.

Show MeSH

Related in: MedlinePlus

Illustration of different Ki-67 assessment methods.A) Analysis of hotspots. B) Analysis of different areas. C) Analysis of only darkly stained nuclei.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4401515&req=5

pone.0123435.g002: Illustration of different Ki-67 assessment methods.A) Analysis of hotspots. B) Analysis of different areas. C) Analysis of only darkly stained nuclei.

Mentions: At 40x magnification on the microscope, in solid tumors, there are approx. 300‐400 cells detectable for the measurements. (Fig 2)


Standardization for Ki-67 assessment in moderately differentiated breast cancer. A retrospective analysis of the SAKK 28/12 study.

Varga Z, Cassoly E, Li Q, Oehlschlegel C, Tapia C, Lehr HA, Klingbiel D, Thürlimann B, Ruhstaller T - PLoS ONE (2015)

Illustration of different Ki-67 assessment methods.A) Analysis of hotspots. B) Analysis of different areas. C) Analysis of only darkly stained nuclei.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4401515&req=5

pone.0123435.g002: Illustration of different Ki-67 assessment methods.A) Analysis of hotspots. B) Analysis of different areas. C) Analysis of only darkly stained nuclei.
Mentions: At 40x magnification on the microscope, in solid tumors, there are approx. 300‐400 cells detectable for the measurements. (Fig 2)

Bottom Line: Fleiss'-kappa-values (14% cut-off) were the highest (moderate) using the original recommendation on light-microscope (Kappa 0.58).Fleiss' kappa values (20% cut-off) were the highest (Kappa 0.48 each) in analyzing hotspots on light-microscopy and digital-analysis.Region analysis and individual review (the original recommendation) on light-microscopy yielded the highest inter-observer reliability.

View Article: PubMed Central - PubMed

Affiliation: Institute of Surgical Pathology, University Hospital Zurich, Zurich, Switzerland.

ABSTRACT

Background: Proliferative activity (Ki-67 Labelling Index) in breast cancer increasingly serves as an additional tool in the decision for or against adjuvant chemotherapy in midrange hormone receptor positive breast cancer. Ki-67 Index has been previously shown to suffer from high inter-observer variability especially in midrange (G2) breast carcinomas. In this study we conducted a systematic approach using different Ki-67 assessments on large tissue sections in order to identify the method with the highest reliability and the lowest variability.

Materials and methods: Five breast pathologists retrospectively analyzed proliferative activity of 50 G2 invasive breast carcinomas using large tissue sections by assessing Ki-67 immunohistochemistry. Ki-67-assessments were done on light microscopy and on digital images following these methods: 1) assessing five regions, 2) assessing only darkly stained nuclei and 3) considering only condensed proliferative areas ('hotspots'). An individual review (the first described assessment from 2008) was also performed. The assessments on light microscopy were done by estimating. All measurements were performed three times. Inter-observer and intra-observer reliabilities were calculated using the approach proposed by Eliasziw et al. Clinical cutoffs (14% and 20%) were tested using Fleiss' Kappa.

Results: There was a good intra-observer reliability in 5 of 7 methods (ICC: 0.76-0.89). The two highest inter-observer reliability was fair to moderate (ICC: 0.71 and 0.74) in 2 methods (region-analysis and individual-review) on light microscopy. Fleiss'-kappa-values (14% cut-off) were the highest (moderate) using the original recommendation on light-microscope (Kappa 0.58). Fleiss' kappa values (20% cut-off) were the highest (Kappa 0.48 each) in analyzing hotspots on light-microscopy and digital-analysis. No methodologies using digital-analysis were superior to the methods on light microscope.

Conclusion: Our results show that all methods on light-microscopy for Ki-67 assessment in large tissue sections resulted in a good intra-observer reliability. Region analysis and individual review (the original recommendation) on light-microscopy yielded the highest inter-observer reliability. These results show slight improvement to previously published data on poor-reproducibility and thus might be a practical-pragmatic way for routine assessment of Ki-67 Index in G2 breast carcinomas.

Show MeSH
Related in: MedlinePlus