UK participation in the UK National External Quality Assessment Scheme for mismatch repair immunohistochemistry: insights from the last decade Authors: Luke AR Farmkiss, Andrew Dodson, Suzanne Parry, Michelle Charmaine James, Ian M Frayling, Mark Johan Arends Published: 2026 ABSTRACT AbstractAimComparison of immunohistochemical detection of mismatch repair (MMR) proteins within tumours, as deficient MMR is important for (1) the detection of Lynch Syndrome, caused by inherited variants affecting the DNA MMR genes MLH1, MSH2, MSH6 and PMS2, (2) aiding MMR gene variant interpretation and (3) deciding on use of immune checkpoint blockade therapy.MethodsThis retrospective analysis compares the performance of different MMR immunohistochemistry (IHC) antibody clones, detection systems and automation IHC platforms using a decade of technical data submitted to the UK National External Quality Assessment Scheme (NEQAS) as part of its MMR EQA programme, with calculation of participants’ final aggregated scores (FAS) for performance comparison.ResultsBetween 2011 and 2022, there were 38 MMR assessment runs, with an average of 44.8 submissions per antigen per assessment run. MMR module participation greatly increased over this decade. Average FAS showed a small non-significant upward trend with the lowest scores (14.6) observed in the first half of the decade, with greater concordance between FAS scores (15.1) in the second half of the decade. For MMR IHC, the antibody clones most frequently used were M1 and ES05 (FAS 15.7 and 15.1) for MLH1, G219-1129 and FE11 (FAS 14.8 and 14.1) for MSH2, EP51 and A16-4 (FAS 15.2 and 15.2) for PMS2 and SP93 and EP49 (FAS 16.1 and 15.6) for MSH6. The most common detection systems were Ventana Optiview, Leica BondMax, Leica Bond Refine, Dako FLEX+ and Ventana UltraView.ConclusionsThe quality of submitted MMR IHC sections has risen, with recent assessment scores showing lower variability, indicating better antibody clone performance, improved detection systems and IHC automation platform technology, allied to increasing technical competence among participants. UK NEQAS provides insight and feedback relating to MMR IHC protocols that most reliably produce high-quality MMR IHC staining to facilitate accurate reporting of the increasingly important tumour MMR status. WHAT IS ALREADY KNOWN ON THIS TOPIC The UK National External Quality Assessment Scheme (NEQAS) experience of mismatch repair (MMR) protein immunohistochemistry (IHC) has not previously been published, so this study provides new information. The UK National External Quality Assessment Scheme (NEQAS) experience of mismatch repair (MMR) protein immunohistochemistry (IHC) has not previously been published, so this study provides new information. WHAT THIS STUDY ADDS UK NEQAS MMR IHC staining quality improved over the decade to 2022, with reduced variability. Increased usage of the better performing primary antibody clones, together with improvements in detection systems and IHC automation platforms contributed to that improved quality. UK NEQAS provided feedback to improve MMR IHC protocols for high-quality MMR immunostaining. UK NEQAS MMR IHC staining quality improved over the decade to 2022, with reduced variability. Increased usage of the better performing primary antibody clones, together with improvements in detection systems and IHC automation platforms contributed to that improved quality. UK NEQAS provided feedback to improve MMR IHC protocols for high-quality MMR immunostaining. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY UK NEQAS plays key roles in monitoring laboratory IHC quality control and helping individual laboratories to improve their MMR IHC with detailed recommendations about use of antibody clones, detection systems, IHC automation platforms and improvements to laboratory protocols. UK NEQAS plays key roles in monitoring laboratory IHC quality control and helping individual laboratories to improve their MMR IHC with detailed recommendations about use of antibody clones, detection systems, IHC automation platforms and improvements to laboratory protocols. Introduction The DNA mismatch repair (MMR) pathway detects and repairs two types of potential mutations that arise during DNA replication: base mismatches and insertion/deletion loops (IDL) that are especially prone to occur in repetitive sequences (eg, microsatellites—repetitive DNA sequences in the normal human genome that are prone to replication slippage during DNA replication, resulting in insertions or deletions, collectively termed InDels). Deficient MMR (dMMR) thus causes accumulation of uncorrected DNA errors, conferring an increased mutation rate and variation in microsatellite lengths (both shorter and longer). Unrepaired IDL within microsatellite sequences change their overall length, resulting in microsatellite instability (MSI).1 2 Detection of dMMR within tumours is critically important for (1) the detection and diagnosis of Lynch Syndrome (LS), caused by inherited variants affecting the DNA MMR genes MLH1, MSH2, MSH6 and PMS2, (2) aiding MMR gene variant interpretation initiatives such as the International Society for Gastrointestinal Tumours Variant Interpretation Committee—now recognised as the ClinGen Variant Classification Expert Panel and the UK’s Cancer Variant Interpretation Group and (3) detection of syndromic or sporadic tumour dMMR is of critical importance in deciding on use of immune checkpoint blockade therapy.111 Although DNA sequencing may be used to diagnose dMMR through determination of a high tumour mutation burden and specific COSMIC signatures (SBS6, SBS14, SBS15, SBS20, SBS21, SBS26, SBS44),12 most detection of tumour dMMR in UK pathology departments is by MMR immunohistochemistry (MMR IHC), which uses specific antibodies to detect the presence or absence of MMR proteins. Absent or abnormal protein expression in cancer cell nuclei indicates underlying MMR deficiency.2 dMMR is present in around 13%–15% of colonic adenocarcinomas.29 Approximately three-quarters of these are sporadic, having arisen from right-sided serrated lesions with hypermethylation of the MLH1 gene promoter, with the remaining one-quarter being due to LS.215 Similarly, 25%–30% of endometrial adenocarcinomas have sporadic dMMR and 3%–4% have dMMR due to LS.57 The Royal College of Pathologists include MMR status as a core dataset item for reporting colorectal cancer resection specimens, and the National Institute for Health and Care Excellence has mandated screening for LS in colorectal cancers by detection of dMMR using either MMR IHC or MSI testing, but in endometrial cancers only using MMR IHC, reflecting its growing clinical importance.10 11 The United Kingdom Accreditation Service (UKAS) ensures medical laboratories comply with international standards (ISO 15189:2022 Medical laboratories—Requirements for quality and competence. International Organization for Standardization, Geneva, Switzerland), by requiring demonstration that laboratory tests are performed to a relevant standard.916 The UK National External Quality Assessment Scheme for Immunocytochemistry and In-Situ Hybridisation (UK NEQAS ICC and ISH) has offered a MMR programme that quality assesses MMR IHC since 2011. UK pathological/medical laboratories must participate within this programme, or otherwise demonstrate MMR IHC testing proficiency, to satisfy UKAS requirements. This retrospective analysis seeks to compare the performance of different MMR IHC antibody clones, detection systems and automation IHC platforms using a decade of submitted technical data. As part of its educational and quality improvement role UK NEQAS ICC and ISH aims to provide insight and feedback relating to the MMR IHC protocols that most reliably produce high-quality MMR IHC staining to facilitate timely and accurate reporting of the increasingly important tumour MMR status. Methods Participating laboratories were invited to perform their in-house MMR IHC protocol on quality-controlled, formalin-fixed, paraffin-embedded tissue sections supplied by UK NEQAS ICC and ISH. Submissions were invited and assessed quarterly, alternating between paired MMR proteins (MLH1/PMS2 and MSH2/MSH6); each MMR protein pair was therefore assessed biannually. Standardisation A cross-section of appendix and sections of both MMR proficient (pMMR) and dMMR colorectal adenocarcinomas were supplied, presectioned, unstained and mounted together on a single glass slide. Sections of the same tissue blocks were also supplied to antibody and platform manufacturers, who then perform MMR IHC to provide reference slides to the scheme. After calibration with manufacturers’ reference slides, four independent assessors scored submissions according to pre-defined standards. Assessors particularly scrutinised the staining gradient in appendiceal mucosal crypt epithelial nuclei (strong staining at the bases of the crypts, fading to almost negative staining at the tops of the crypts). Assessors expected retained MMR protein expression within pMMR tumour cell nuclei and internal control stromal cells and lymphocytes, but MMR protein expression was only within the internal control stromal cells and lymphocytes of dMMR tumours with loss of staining in their tumour cell nuclei. Scoring Assessors’ scores ranged from 1 to 5. Scores of 1–2 reflected unacceptable, non-diagnostic, staining quality and were deemed a test failure. Scores of 3 were used for borderline staining that was technically suboptimal but produced clinically appropriate results. Scores of 4 were given for clinically acceptable staining quality with only minor technical issues. Scores of 5 were provided for optimal staining quality. Submissions with mixed scores (ie, both failing (1–2) and passing scores (3–5)) were re-reviewed and discussed until a consensus opinion among assessors was reached. Each submission’s final aggregated score (FAS) was the sum of the individual assessor scores. FAS were categorised into three outcomes: unacceptable (FAS score: 0–9), borderline acceptable (FAS score: 10–12) and acceptable (FAS score: 13–20). Unacceptable submissions were considered to have failed quality assessment. Analysis FAS was compared with primary antibody clone, IHC detection system and automated staining platform data, which were curated and analysed in Excel (Microsoft) and Prism. Statistical comparisons of FAS data used the Welch’s t-test (two-tailed for unequal sample sizes and variances). Results Between 2011 and 2022, there were 38 assessment runs, labelled Run 95 to Run 135, with 3405 submissions from 100 unique UK and UK Crown Dominion (Channel Islands; Isle of Man) participants. An average of 44.8 submissions per antigen were submitted per assessment run (Run 95; 40 submissions (20 MLH1, 20 PMS2) vs Run 135; 175 submissions (87 MSH2, 88 MSH6)) (table 1). Table 1A table showing the average number, SD and total of submissions per assessment run (left side), and the average number, SD and total number of unacceptable submissions (with a FAS of 0–9) per assessment run (right side), between 2011 and 2022Average submissions per runAverage unacceptable submissions per runAntigenAverageSDTotalAverageSDTotalMLH144.411.38443.83.657PMS244.411.38443.52.360MSH245.111.18572.91.644MSH645.311.18603.62.765Total44.811.234053.52.6226FAS, final aggregated score. A table showing the average number, SD and total of submissions per assessment run (left side), and the average number, SD and total number of unacceptable submissions (with a FAS of 0–9) per assessment run (right side), between 2011 and 2022 FAS, final aggregated score. The average FAS converged around the mean across assessment runs and showed a small non-significant upward trend (figure 1). Hence, lowest scores were observed in the first half of the observed period (run 95–117 Average FAS 14.6, StdDev 0.89, Var 0.79, IQR 1.2) with run 103 showing the lowest average FAS at 12.6. There were no significant outliers between assessment runs 95–117. Simple linear regression of total average FAS across all submissions between 2011–2022. FAS, final aggregated score. Greater concordance between scores was observed in the latter half of the observed period (run 118–135) (Average FAS: 15.1, SD 0.54, Var 0.29, IQR 0.59). Between run 118–135, run 120 represented a significant outlier (average FAS: 16.6, Q3+1.5 IQR=16.2). Assessment run 130 showed an average FAS of 14.1 and was close to the outlier threshold (Q1–1.5 IQR=13.8) (figure 1). Borderline acceptable or better scores were achieved in 93.4% of submissions (n=3179). Unacceptable scores were found in 226 submissions (6.6% of total submissions; 6.8% of MLH1 submissions; 7.1% of PMS2 submissions; 5.1% of MSH2 submissions; 7.6% of MSH6 submissions). The average number of unacceptable submissions per assessment run per antigen ranged between 2.9 (MSH2) and 3.8 (MLH1). Unacceptable submission rate per antigen ranged from 5.1% (MSH2) to 7.6% (MSH6) (table 1). Clones MLH1 ES05 has remained the most submitted MLH1 antibody clone throughout the study period and accounts for the highest proportion of submissions for any clone over all antibody clones (56.5% of MLH1 submissions). ES05 has an average FAS of 15.1 whereas the best performing MLH1 clone, M1, has a similar but slightly and significantly better average FAS of 15.7 (Welch’s t-test (two-tailed for unequal sample sizes and variances), p=0.0038, online supplemental table 1). Unacceptable submissions of ES05 and M1 occurred in 36 (7.5%, HR: 1.1) and 10 (3.2%, HR: 0.5) submissions respectively between 2011 and 2022 (figure 2A, table 2). In assessment run 130, ES05 scored an average FAS of 12.1 over 37 submissions which is a significant singular outlier. Graphs depicting the number of submissions of selected clones over the study period for antibodies against (A) MLH1, (B) PMS2, (C) MSH2 and (D) MSH6. Table 2Popular antigen clones’ performances tabulatedAntigenCloneCountProportion (%)FASCount of failedFailed (%)HRMLH1M131537.315.7103.20.5ES0547756.515.1367.51.1Other526.212.61121.23.1Subtotal84414.5576.8PMS2EP5133239.315.2164.80.7EPR394719923.614.7157.51.1A16-423728.115.2156.30.9Other769.013.71418.42.6Subtotal84414.7607.1MSH2G219-112935541.414.8154.20.8FE1132537.914.1185.51.179H1113515.815.385.91.2Other424.914.537.11.4Subtotal85714.7445.1MSH6EP4942249.115.6184.30.6SP9316218.816.131.90.34416619.313.42012.01.6PU29505.812.01938.05.0Other607.014.858.31.1Subtotal86014.3657.6 Total3405226Count: the number of submissions per clone between 2011 and 2022. Proportion (%): the proportion of submissions per clone expressed as a percentage of the subtotal of submissions received for the relevant antigen. FAS: the average FAS of submissions per clone between 2011–2022. Count of failed: the number of submissions with FAS <10 per clone. Failed (%): the proportion of submissions with a FAS <10 per clone, expressed as a percentage of the total number of submissions received for that clone. HR: ratio of the failure rate of a clone divided by the overall failure rate for the antigen. Clones with fewer than 40 submissions were aggregated and designated ‘other’.FAS, final aggregated score. Popular antigen clones’ performances tabulated Count: the number of submissions per clone between 2011 and 2022. Proportion (%): the proportion of submissions per clone expressed as a percentage of the subtotal of submissions received for the relevant antigen. FAS: the average FAS of submissions per clone between 2011–2022. Count of failed: the number of submissions with FAS <10 per clone. Failed (%): the proportion of submissions with a FAS <10 per clone, expressed as a percentage of the total number of submissions received for that clone. HR: ratio of the failure rate of a clone divided by the overall failure rate for the antigen. Clones with fewer than 40 submissions were aggregated and designated ‘other’. FAS, final aggregated score. PMS2 The most submitted PMS2 antibody clone is EP51, although A16-4 has shown recent popularity. Both EP51 and A16-4 show an average FAS of 15.2 (not significantly different, online supplemental table 1), in contrast to 14.7 for EPR3947. EP51 has a lower rate of unacceptable submissions (4.8%, HR 0.7) than A16-4 (6.8%, 0.9). There was no significant increase in the average FAS for EP51 or A16-4 over assessment runs (figure 2B, table 2). MSH2 G219-1129 and FE11 account for 41.4% vs 37.9% of MSH2 submissions respectively and share a similar average FAS (14.8, G219-1129 vs 14.1, FE11, but not significantly different, online supplemental table 1) and a similar rate of unacceptable submissions (4.2%, HR: 0.8, G219-1129 vs 5.5%, HR: 1.1, FE11). Clone 79H11 shows the highest average FAS among MSH2 clones (15.3) (figure 2C, table 2). MSH6 EP49 is the most submitted MSH6 clone overall (49.1%), although in the latest run (135), EP49 accounts for 44.9% of submissions (n=39) whereas SP93 has increased to 40.9% of MSH6 submissions. SP93 shows a greater average FAS than EP49 (16.1 vs 15.6, that is significantly different, online supplemental table 1) and a lower rate of unacceptable submissions (1.9%, HR: 0.3 vs 4.3%, HR: 0.6). EPR3945, 44 and PU29 account for 61.5% of unacceptable MSH6 and are only rarely submitted in the most recent assessment runs (figure 2D, table 2). Statistical comparison of the major antibody clones’ FAS performances was by Welch’s t-tests, which indicated significant differences between the FAS data for several of the antibody clones compared for each antigen (MLH1, MSH2, MSH6 and PMS2) (online supplemental table 1). This analysis was informative, providing some guidance for use of the best-performing antibody clones: M1 or ES05 for MLH1, FE11 or G219-1129 for MSH2, SP93 or EP49 for MSH6, and EP51 or A16-4 for PMS2 in this dataset. Detection systems Data on participants’ implemented detection systems were available in 3247 (95.4%) submissions (table 3). The most common detection systems were Ventana Optiview (n=1221, 37.6%); Leica BondMax Refine KIT (n=892, 27.5%); Dako FLEX+kit (n=542, 15.8%); Leica Bond Refine (n=273, 8.4%) and Ventana UltraView DAB (n=158, 4.9%). Average FAS ranged from 13.9 (Ventana UltraView DAB) to 15.8 (Dako FLEX+kit) across the five most submitted detection systems. The BioGenex SS Polymer-HRP scored the highest average FAS across all detection systems (15.9), but this detection system was only implemented in 0.55% of submissions (n=22). Table 3The total number of submissions and the number of failed submissions per detection systemDetection systemsDetection systemn totaln failed% FailedHR% of Failed cases (/226)Leica BondMax Refine KIT892819.10.635.8Ventana Optiview1221604.90.326.5Dako FLEX+kit542275.00.311.9Ventana UltraView DAB1582012.70.88.8Leica Bond Refine273145.10.36.2Envision28621.41.42.7Dako FLEX+kit12541.72.72.2As per kit/assay5647.10.51.8Not specified21314.30.91.3Ventana iView Kit6350.03.21.3Dako EnVision FLEX (Omnis)18211.10.70.9Bond polymer refine DS98002015.00.30.4Totals (average)3247226(15.6)% Failed: the proportion of failed submissions as a percentage of the total submissions received for the detection system. HR: a ratio of % Failed divided by the average % Failed (15.6%). % of Failed cases (/226): the proportion of each detection systems failed submissions as a percentage of the total number of failed cases. The total number of submissions and the number of failed submissions per detection system % Failed: the proportion of failed submissions as a percentage of the total submissions received for the detection system. HR: a ratio of % Failed divided by the average % Failed (15.6%). % of Failed cases (/226): the proportion of each detection systems failed submissions as a percentage of the total number of failed cases. The five most common detection systems accounted for 95% of total submissions with available data and showed an average unacceptable submission rate of 7.4% (range 5%–13%). Ventana Optiview showed the lowest failure rate (5%) in this group, while being the most submitted detection system. Ventana UltraView DAB showed the highest failure rate (12.7%) in this group, correlating with the lower average FAS seen in submissions implementing this detection system (figure 3). Average FAS of each antigen on the five most submitted detection systems. FAS, final aggregated score. The failure rate among detection systems ranged from 4.5% to 100% (average failure rate: 24.5%). Two significant outliers were removed (Leica Bond X, n=1, failure rate, 100%; and Dako Autostainer Link 48 n=1, failure rate, 100%). The adjusted failure rate among detection systems ranged from 4.5% to 25% (adjusted average failure rate: 10.8%). The most common automation system among unacceptable submissions was the Leica Bond-III (76 failed cases; 7.5% failure rate; HR: 0.69). Examples of strong or very strong (acceptable) MMR IHC staining and weak or very weak (unacceptable) MMR IHC staining are shown in figure 4. UK NEQAS mismatch repair (MMR) protein module immunohistochemistry (IHC) photomicrographs showing some examples of very good and unacceptably weak MMR IHC staining. (A) Very good MLH1 IHC staining of control normal appendix, showing the typical gradient of strong staining at the base of the crypts fading to weaker staining in the upper crypts with negative staining in the epithelial table in contact with the appendix lumen. There is moderate staining of the lamina propria cells and strong lymphocyte staining of the lymphoid follicles (lower left) (Magnification ×5). (B) Weak and variable MLH1 IHC staining of a control MMR-proficient colorectal cancer, showing variable strengths of staining in the malignant glands with weak staining in the tumour stroma microenvironment (Magnification ×5). (C) Very good strong MSH2 IHC staining of most of the stromal cells in a test MSH2-deficient colorectal cancer, with negative staining in the malignant clusters and glands (Magnification ×10). (D) Very good strong MSH2 IHC staining of both malignant clusters and glands in a control MMR-proficient colorectal cancer (Magnification ×10). (E) Very good strong MSH6 IHC staining of most of the stromal cells in a test MSH6-deficient colorectal cancer, with negative staining in the malignant islands and glands (Magnification ×10). (F) Weak and variable MSH6 IHC staining of a control MMR-proficient colorectal cancer, showing variable but mostly weak or very weak staining in the malignant glands and very weak to negative staining in the tumour stroma microenvironment (Magnification ×10). (G) Very good strong PMS2 IHC staining of control normal appendix, showing the typical gradient of strong staining at the base of the crypts fading to weaker staining in the upper crypts with negative staining in the epithelial table in contact with the appendix lumen (Magnification ×10). (H) Very good strong PMS2 IHC staining of most of the stromal cells in a test PMS2-deficient colorectal cancer, with negative staining in the malignant glands (Magnification ×10). NEQAS, National External Quality Assessment Scheme. Automation Data on automation IHC platform were available in 3403 (99.9%) submissions. The most common automation systems among submissions were the Ventana Benchmark ULTRA (n=1083, 32.6%) and the Leica Bond-III (n=1017, 30%). The Ventana Benchmark ULTRA and Leica Bond-III scored an average FAS of 15.3 and 14.7, respectively. The Dako Autostainer Link 48 scored the highest average FAS (15.96) and accounted for 11.7% of submissions (n=399). Discussion Scheme growth Increasing clinical relevance of tumour MMR status and implementation of national LS screening has driven centres across the UK and UK CD to add MMR IHC to their laboratories’ repertoire. These laboratories must demonstrate competency in testing to conform with international standards, usually via EQA participation. This has caused the UK NEQAS MMR IHC scheme to grow 4.4-fold over the decade. Further growth is still anticipated as both international and domestic interest in tumour MMR status continues to develop. Given the finite number of centres within the UK’s National Health Service (NHS), it is feasible that scheme growth may reach a zenith as the implementation of LS screening settles across the UK.1114 Newer technologies may also become available, which could challenge the screening paradigm and cause decreased scheme participation if the need for MMR IHC wanes as other methods increase, such as tumour genome/exome sequencing.121519 Participant feedback The overall rate of unacceptable submissions was 6.6% (n=226) over ten years. Assessors provided technical feedback to all of these participants, as well as those who received borderline scores, to help improve IHC staining quality. This is a critical function of the MMR IHC UK NEQAS and facilitates national service improvement. Participants who submit repeated unacceptable material are provided further support, but are unable to continue performing MMR IHC diagnostically until there is a satisfactory submission to a subsequent EQA assessment run. This helps to ensure patient safety. The role of providing feedback to participants played an important part in providing detailed advice on how to improve staining quality over the decade. The most common feedback given was that there was weak demonstration of antigen and/or excessive background staining. Participants were advised to review their protocols, with specific UK NEQAS advice provided on antigen retrieval times and antibody dilutions, which are frequent causes of underperformance. Submission improvement The reagents and hardware available to participants have improved over the last ten years. Participants have gained greater familiarity and increasing competency with MMR IHC testing over the course of the MMR UK NEQAS scheme’s lifetime. These factors are reflected in the upward trend demonstrated in figure 1 and the lesser degree of variability in FAS scores in later assessment runs. Scheme limitations There is interdependency of the variables investigated, including antibody clones, IHC platforms, detection systems, and protocol variables, and interactions between these factors could influence staining performance. MMR IHC is an interpretative test with subjective limitations. We have tried to control for these by seeking the independent opinion of four individual expert assessors drawn from a range of technical, scientific and clinical backgrounds. Assessors submit scores as discretely as possible, but discrepant scores (where one assessor has failed a case while another one has passed it) are flagged to all for rediscussion and reassessment, therefore assessors are ultimately not blinded to one another’s scores, which might lead to interpretative bias. If assessors do not submit scores discretely, there is also a further risk of conformity bias. It is likely that rediscussion of discrepant cases fosters greater quality feedback to participants than if assessors were formally blinded and potentially confusing discrepant results returned. Any performance analysis of participants’ MMR IHC protocols is also confounded by the myriad variables inherent with the scheme such as tissue quality, fixation and sample preparation, including sectioning methods and mounting. By providing participants with preprepared UK NEQAS control tissue sections there should be mitigation of these factors. However, there are further confounding variables including participant antigen retrieval times, counterstaining methods and antibody dilution ratios. The data is incomplete on these variables, limiting full analysis. Further, for data capture from the participating laboratories, their NHS/private activities were not formally captured and the diagnostic-volume data were not available. Clones MLH1 Both the M1 and ES05 clone have shown increasing popularity over the course of the scheme since both showed strong performance with average FAS of 15.7 and 15.1, respectively. Removing the underperforming outlier results for ES05 from run 130 improves the average FAS of ES05 to 15.4. These two clones appear to have become more popular than others presumably from laboratory shared experiences of the technical aspects. PMS2 The increased use of A16-4 is proportional and most likely directly related to the decreased submissions received for EPR3947, possibly relating to a clear difference in laboratory performance of these clones with one replacing the other. MSH2 Recent assessment runs have seen the increasing popularity of 79H11, which shows the highest average FAS among MSH2 clones. Its increased adoption may be related to its observed improved laboratory performance. MSH6 SP93 has gained recent popularity after introduction sometime after assessment run 121. Increased adoption of this clone may be related to its performance over the more widespread EP49 clone. Clone 44 use has dropped off significantly since the introduction of SP93, which is appropriate, as Clone 44 performed poorly in the laboratory. Conclusions Participation in the UK NEQAS Scheme MMR module has greatly increased over the last decade, which is reflective of the increasing clinical importance of MMR IHC. The quality of MMR IHC submissions has improved, and more recent assessment scores have shown lower levels of variability, indicating improved antibody clone performance, better detection systems and IHC automation platform technology use, with increasing technical competence among participants, although there is interdependency of these factors. Certain antibody clones have come to dominate MMR IHC, mostly with good performance shown by acceptable FAS scores, although newcomers do occasionally break through. Antibody clones with higher FAS performance scores and most frequently used are M1 and ES05 (FAS 15.7 and 15.1) for MLH1, G219-1129 and FE11 (FAS 14.8 and 14.1) for MSH2, EP51 and A16-4 (FAS 15.2 and 15.2) for PMS2, and SP93 and EP49 (FAS 16.1 and 15.6) for MSH6—and these represent current UK NEQAS recommendations for MMR IHC antibody clone use. The most commonly used detection systems are Ventana Optiview, Leica BondMax, Leica Bond Refine, Dako FLEX+and Ventana UltraView. The Dako Flex+kit demonstrated the greatest FAS among detection systems and is recommended. The Dako Autostainer Link 48 scored the highest FAS among IHC automation platform systems and is recommended, despite only accounting for 11.7% of submissions. No antibody clone, detection system or IHC automation platform system is consistently associated with unacceptable scores. Overall, the UK NEQAS MMR module data illustrate improving laboratory performance over the decade of data analysis and provide a basis for UK NEQAS advice and recommendations. Supplementary material