Description of Event or Problem · 0
SITUATION: INACCURATE INFORMATION ON THE QUALITY CONTROL AND MAINTENANCE PLAN IN THE UPDATED EDUCATIONAL PAMPHLET ON FOLR1 TESTING USING ROCHE-VENTANA'S RXDX PLATFORM (COMPANION DIAGNOSTIC FOR ELAHERE) LED TO AT LEAST ONE FALSE NEGATIVE TEST RESULT. BACKGROUND: FOLR1 TEST USING ROCHE-VENTANA'S RXDX PLATFORM IS THE COMPANION DIAGNOSTIC (CDX) FOR THE THERAPEUTIC PRODUCT ELAHERE (MIRVETUXIMAB SORAVTANSINE-GYNX) AND IS READ AS POSITIVE WHEN 1) THE TEST IS SET FOR APPROPRIATE ANALYTICAL SENSITIVITY AND 2) >75% OF TUMOR CELLS ARE POSITIVE AND 3) THE POSITIVE STAINING IS MEMBRANOUS AND 4) THE INTENSITY OF THE STAINING IS MODERATE/STRONG. DURING THE TRIAL, ANALYTICAL SENSITIVITY (SLIDE INTENSITY REFERENCE (SIR)) WAS ESTABLISHED USING NORMAL FALLOPIAN TUBE TISSUE WHICH REQUIRED 2+ MEMBRANOUS STAINING OF AT LEAST 10 TUBAL EPITHELIAL CELLS, WITH STROMAL CELLS EXHIBITING NO (ZERO) STAINING, AND THE APICAL/LUMINAL MEMBRANE DISPLAYING STRONG (3+) STAINING. OF NOTE, 2+ IS NOT A PRECISE VALUE, AND REPRESENTS A VALUE BETWEEN 1+ (WEAK) AND 3+(STRONG). PATHOLOGISTS WERE THUS INSTRUCTED TO CALIBRATE THEIR 2+ INTENSITY READING BY ALIGNING THEMSELVES TO THE SIR BEFORE EVALUATING THE TEST CASE. HOWEVER, THESE REQUIREMENTS WERE REVISED BASED ON THE DEMONSTRATED EQUIVALENCY OF PERFORMANCE IN A SIMULATED BRIDGING STUDY, BETWEEN SIR AND NON-SIR ARMS. THE UPDATED PROTOCOL SHIFTED TO REQUIRING ONLY A POSITIVE CONTROL WITH NO GUIDANCE ON ASSESSING ANALYTICAL SENSITIVITY. NORMAL FALLOPIAN TUBE WAS CITED AS AN EXAMPLE OF A POSITIVE CONTROL. THE USE OF TUMOR TISSUE WITH INSUFFICIENT ANALYTICAL SENSITIVITY AS A POSITIVE CONTROL LED TO A LEFT-SHIFTED TEST. THE INITIALLY USED CONTROL SHOWED 2+ STAINING. HOWEVER, UPON CALIBRATING TO A SIR-FALLOPIAN TUBE CONTROL, IT SHOWED PREDOMINANT 3+ STAINING. RETESTING OF THE INCIDENT CASE, LED TO A RESULT REVISION FROM NEGATIVE (50-60%, 2+) TO POSITIVE (80% , 2+ AND 3+ INTENSITY). ASSESSMENT: MY ASSESSMENT FOR THIS FAILURE ARE 1. LACK OF INFORMATION ABOUT SETTING ANALYTICAL SENSITIVITY FOR THE TEST IN THE NEW PAMPHLET. BEST PRACTICE FOR IMMUNOHISTOCHEMISTRY IS TO SET ANALYTICAL SENSITIVITY FOR THE LOWER LIMIT OF DETECTION (LLOD) (THE EQUIVALENT OF THE SIR CONTROL). HOWEVER DEFINING THE LLOD FOR NEW STAINS IS DIFFICULT UNTIL SIGNIFICANT EXPERIENCE HAS ACCUMULATED WITH THE STAIN. THE ONLY LLOD CONTROL USED BY VENTANA IN THEIR TRIAL WAS NORMAL FALLOPIAN TUBE. THE LAB USED TUMORAL TISSUE AS A 'POSITIVE' CONTROL IN ACCORDANCE WITH THE UPDATED PAMPHLET INSTRUCTION. A LAB CAN ONLY IDENTIFY AN ALTERNATIVE LLOD CONTROL (SECONDARY REFERENCE STANDARD) AFTER APPROPRIATELY CALIBRATING IT TO THE KNOWN LLOD CONTROL/PRIMARY REFERENCE STANDARD. TUMOR TISSUE IS NOT IDEAL AS AN LLOD CONTROL. 2. LACK OF INFORMATION FROM VENTANA ON A PROCESS FOR ENSURING THE PROFICIENCY OF PATHOLOGISTS ASSIGNED TO REPORT THE TEST ('ALL PATHOLOGISTS CAN INTERPRET ANY IHC'). PATHOLOGISTS INTERPRETED TEST RESULTS WITHOUT FULL KNOWLEDGE OF THE MEASURES OF PRECISION AND ACCURACY OF THE TEST. THE CONTROL WAS DEEMED ADEQUATE WHEN IT WAS NOT. 3. FAILURE TO RECOGNIZE THAT A PROCESS CHANGE LED TO THE NON-DELIVERY OF POSITIVE CONTROLS TO PATHOLOGISTS FOR THEIR REVIEW. PATHOLOGISTS SEEMINGLY INTERPRETED THE TEST WITHOUT CONFIRMING CONTROL PERFORMANCE. 4. THE LACK OF INFORMATION FROM VENTANA ON A PROCESS FOR ONGOING PERFORMANCE MONITORING OF THIS TEST LED TO THE LATE DETECTION OF POSSIBLE ASSAY DRIFT. RECOMMENDATIONS/ SUGGESTIONS: 1.VENTANA SHOULD REVERT TO THE ORIGINAL INSTRUCTIONS WITH A.CLEAR STATEMENT THAT 'A WEAK POSITIVE CONTROL IS RECOMMENDED BEST PRACTICE' FOR THIS TEST. B.EXPLICITLY MANDATE ESTABLISHING ANALYTICAL SENSITIVITY USING A NORMAL FALLOPIAN TUBE. A SECONDARY REFERENCE STANDARD MAY BE IDENTIFIED AFTER CALIBRATING TO THIS STANDARD. C. EXPLICITLY REQUIRE PATHOLOGISTS TO CALIBRATE THE ASSESSMENT OF 2+ BY FIRST REVIEWING THE SIR/LLOD CONTROL. 2.VENTANA SHOULD INFORM ON A PROFICIENCY ASSESSMENT PLAN FOR PATHOLOGISTS. RELATEDLY, THE FDA SHOULD REQUIRE THAT ALL CDX SUBMISSIONS INCLUDE A READER PROFICIENCY ASSESSMENT PLAN, GIVEN THAT CDX'S ARE HIGH-PENALTY TESTS. HIGH INTERLABORATORY AND INTER-PATHOLOGIST PERFORMANCE WAS NOT ACHIEVED FOR HER2 IHC UNTIL EXTERNAL QUALITY ASSURANCE PROFICIENCY TESTING PLANS WERE (CAP). ALTHOUGH THE JOINT COMMISSION REQUIRES A FOCUSED PROFESSIONAL PRACTICE EVALUATION BEFORE PRACTITIONERS EXPAND INTO A NEW PRACTICE DOMAIN, MOST LABS DO NOT CONSIDER A 'NEW IHC TEST' A NEW 'PRACTICE DOMAIN.' A STRAIGHTFORWARD ONLINE MULTIPLE-CHOICE TEST, WHERE PATHOLOGISTS DEMONSTRATE THEIR KNOWLEDGE OF THE STAINING WOULD ENSURE INTERPATHOLOGIST PRECISION AND ACCURACY OF TEST RESULTS. THIS WILL ALSO FACILITATE POST-MARKET SURVEILLANCE AND ASSESSMENT OF TEST PERFORMANCE. 3.VENTANA SHOULD PROVIDE, AND FDA SHOULD REQUIRE THAT ALL CDX INSTRUCTION PAMPHLETS PROVIDE AN OUTLINE FOR A CONTINUED QUALITY MONITORING PLAN. INDIVIDUAL QUALITY CONTROL PLANS ARE NOT USUAL IN ANATOMIC PATHOLOGY (AP). 4.VENTANA SHOULD PROVIDE, AND FDA SHOULD REQUIRE ALL CDX INSTRUCTION PAMPHLETS TO INCLUDE INFORMATION FOR REQUIRED REPORTING OF TEST-PERFORMANCE FAILURES TO THE MANUFACTURER AND FDA BY USER LABS. AP LABS ARE NOT AWARE OF THESE REQUIREMENTS. 5. 5.FDA SHOULD REQUIRE GREATER DETAIL FOR SIMULATED BRIDGING STUDIES. I COULD NOT ASSESS IF THE RESULTS WERE APPLICABLE TO SIR/NON-SIR ARMS IN A NEW LAB THAT IS HAD NOT ALREADY OPTIMIZED PERFORMANCE.