As a company-wide quality metrics, ISR pass/fail and observations are continuously documented and updated on a quarterly basis.

Similar documents
Treatment Spring Late Summer Fall Mean = 1.33 Mean = 4.88 Mean = 3.

An Undergraduate Curriculum Evaluation with the Analytic Hierarchy Process

Test Management using Telelogic DOORS. Francisco López Telelogic DOORS Specialist

Techniques for Requirements Gathering and Definition. Kristian Persson Principal Product Specialist

Economics Letters 65 (1999) macroeconomists. a b, Ruth A. Judson, Ann L. Owen. Received 11 December 1998; accepted 12 May 1999

AN ANALYTICAL HIERARCHY PROCESS METHODOLOGY TO EVALUATE IT SOLUTIONS FOR ORGANIZATIONS

ClearPeaks Customer Care Guide. Business as Usual (BaU) Services Peace of mind for your BI Investment

Rate and Activation Energy of the Iodination of Acetone

JaERM Software-as-a-Solution Package

DlNBVRGH + Sickness Absence Monitoring Report. Executive of the Council. Purpose of report

Reasoning to Solve Equations and Inequalities

Recognition Scheme Forensic Science Content Within Educational Programmes

LINEAR TRANSFORMATIONS AND THEIR REPRESENTING MATRICES

Vendor Rating for Service Desk Selection

Example 27.1 Draw a Venn diagram to show the relationship between counting numbers, whole numbers, integers, and rational numbers.

2015 EDITION. AVMA Report on Veterinary Compensation

Polynomial Functions. Polynomial functions in one variable can be written in expanded form as ( )

Operations with Polynomials

The LENA TM Language Environment Analysis System:

Math 135 Circles and Completing the Square Examples

Enterprise Risk Management Software Buyer s Guide

Project 6 Aircraft static stability and control

E-Commerce Comparison

Application Bundles & Data Plans

Lecture 3 Gaussian Probability Distribution

SyGEMe: Integrated Municipal Facilities Management of Water Ressources Swiss Geoscience Meeting, Neuchâtel, 21 novembre 2009 k

Engineer-to-Engineer Note

Why is the NSW prison population falling?

Section 5-4 Trigonometric Functions

4.11 Inner Product Spaces

Greg Pope, Analytics and Psychometrics Manager 11am-12:30pm, Monday March 15

Utilization of Smoking Cessation Benefits in Medicaid Managed Care,

Learner-oriented distance education supporting service system model and applied research

10.6 Applications of Quadratic Equations

Corporate Compliance vs. Enterprise-Wide Risk Management

Decision Rule Extraction from Trained Neural Networks Using Rough Sets

STATE OF MONTANA Developomental Disabilities Program Comprehensive Evaluation Hi-Line Home Programs, Inc Adult Services

** Dpt. Chemical Engineering, Kasetsart University, Bangkok 10900, Thailand

9 CONTINUOUS DISTRIBUTIONS

Trade liberalization and per capita income convergence: a difference-in-differences analysis

Engineer-to-Engineer Note

How To Study The Effects Of Music Composition On Children

FDIC Study of Bank Overdraft Programs


Prescriptive Program Rebate Application

Graphs on Logarithmic and Semilogarithmic Paper

Econ 4721 Money and Banking Problem Set 2 Answer Key

Data quality issues for accounting information systems implementation: Systems, stakeholders, and organizational factors

Characteristics of Applicants Who Obtain Interviews at Orthodontic Postgraduate Programs

Helicopter Theme and Variations

improved (2, 3, 4, 5, 7, 8, 9, 10, 11). Some of the

Performance analysis model for big data applications in cloud computing

Factoring Polynomials

Introducing Kashef for Application Monitoring

THERMAL EXPANSION OF TUNGSTEN

STATUS OF LAND-BASED WIND ENERGY DEVELOPMENT IN GERMANY

Multiple Testing in a Two-Stage Adaptive Design With Combination Tests Controlling FDR

Unit 29: Inference for Two-Way Tables

Quick Reference Guide: One-time Account Update

National greenhouse gas inventory data for the period

MAX. As an increasingly larger share of Medicaid enrollees MEDICAID POLICY BRIEF

GAO IRS AUDIT RATES. Rate for Individual Taxpayers Has Declined But Effect on Compliance Is Unknown

RP-HPLC method development and validation for estimation of rivaroxaban in pharmaceutical dosage forms

Small Business Cloud Services

Commercial Cooling Rebate Application

Cost Functions for Assessment of Vehicle Dynamics

Appendix D: Completing the Square and the Quadratic Formula. In Appendix A, two special cases of expanding brackets were considered:

The usability study details initial testing of the GIMCF-AHP prototype in a practical MADM task or environment.

Experiment 6: Friction

COMPLEX FRACTIONS. section. Simplifying Complex Fractions

Hillsborough Township Public Schools Mathematics Department Computer Programming 1

Small Businesses Decisions to Offer Health Insurance to Employees

E-Mentoring and Information Systems Effectiveness Models: a useful nexus for evaluation in the small business context.

Distributions. (corresponding to the cumulative distribution function for the discrete case).

Space Vector Pulse Width Modulation Based Induction Motor with V/F Control

Integration. 148 Chapter 7 Integration

Numeracy across the Curriculum in Key Stages 3 and 4. Helpful advice and suggested resources from the Leicestershire Secondary Mathematics Team

Active & Retiree Plan: Trustees of the Milwaukee Roofers Health Fund Coverage Period: 06/01/ /31/2016 Summary of Benefits and Coverage:

The Definite Integral

Assessing authentically in the Graduate Diploma of Education

Modeling POMDPs for Generating and Simulating Stock Investment Policies

MATH 150 HOMEWORK 4 SOLUTIONS

ARTICLE IN PRESS. i n t e r n a t i o n a l j o u r n a l o f m e d i c a l i n f o r m a t i c s x x x ( ) xxx xxx

Basic Analysis of Autarky and Free Trade Models

PROF. BOYAN KOSTADINOV NEW YORK CITY COLLEGE OF TECHNOLOGY, CUNY

TITLE THE PRINCIPLES OF COIN-TAP METHOD OF NON-DESTRUCTIVE TESTING

Contextualizing NSSE Effect Sizes: Empirical Analysis and Interpretation of Benchmark Comparisons

Health Information Systems: evaluation and performance of a Help Desk

D T IC. 052 Research Laboratory D A 0 05

APPLICATION OF TAGUCHI EXPERIMENTAL DESIGN FOR PROCESS OPTIMIZATION OF TABLET COMPRESSION MACHINES AT HLL LIFECARE LIMITED, INDIA

Quality Evaluation of Entrepreneur Education on Graduate Students Based on AHP-fuzzy Comprehensive Evaluation Approach ZhongXiaojun 1, WangYunfeng 2

Immunoglobulins in Umbilical Cord Plasma

Portfolio approach to information technology security resource allocation decisions

Understanding Basic Analog Ideal Op Amps

Project Recovery. . It Can Be Done

Lump-Sum Distributions at Job Change, p. 2

Rotating DC Motors Part II

COMPARISON OF SOME METHODS TO FIT A MULTIPLICATIVE TARIFF STRUCTURE TO OBSERVED RISK DATA BY B. AJNE. Skandza, Stockholm ABSTRACT

Uplift Capacity of K-Series Open Web Steel Joist Seats. Florida, Gainesville, FL 32611;

Transcription:

Appliction Note #2009-LC/MS/MS-001 >>> Incurred Smple Renlysis Introduction The Incurred Smple Renlysis (ISR) progrm hs been implemented in most bionlyticl lbortories s n dditionl mesure to evlute the relibility of the dt nd ssy performnce. A filed ISR requires n investigtion to determine why the ssy is not performing t the sme level during smple nlysis s compred to the initil method vlidtion runs. On the other hnd, successful ISR ssessment helps to improve confidence in the relibility of dt. Although QPS hs been performing ISR on some smples since 2003, we implemented compny-wide SOP on ISR s n dditionl mesure for qulity on Jnury 21, 2008. This ppliction note documents some of the ISR findings for LC/MS/MS ssy over 14 month period. Tbles 1 nd 2 re the most current metrics. Tble 1 represents n overview of the percentge ISR pss rte by the # of smples nd the # of studies within the first 14 months when ISR ws implemented t QPS compny-wide. Tble 2 is more detiled metrics ech br represents one study, nd the percent devition from the originl dt for ech ISR smple is grouped together into three distinct bnds within ech study. Tble 2 gives us cler visul metric of how mny smples pss ISR nd how close the ISR dt is to the originl dt. Two cse studies re lso presented to discuss some of the more interesting ISR observtions. As compny-wide qulity metrics, ISR pss/fil nd observtions re continuously documented nd updted on qurterly bsis. Tble 1. LC/MS/MS ISR Pss Rte (1/21/2008 4/3/2009) 100 % of ISR Pssed 95 90 85 91.2 Out of 4059 Smples 92.2 Out of 84 Studies 80 % By the # of Smples % By the # of Studies LC/MS/MS Studies Tble 2. Stck Plot of % Devition from Originl Vlue (1/21/2008 4/3/2009) 100% % Smples In Study 66.67% 75% 50% 25% % >50% % >20% <50% % <20% 0% 1 11 21 31 41 51 61 71 81 Study #

Cse Study 1 This is 4-in-1, prent nd metbolites, dog plsm ssy developed to support niml sfety studies. A stble isotope lbeled (SIL) ws used s n internl stndrd (IS) for the prent compound, A, nd n nlogue ws used s IS for the three metbolites (B, C, nd D). The ccurcy nd precision for ll four nlytes during method vlidtion rnged from -11.8% to 13.4% nd 1.4% to 14.9%, respectively. ISR Summry Compound # of ISR Smples # of Smples Within +/- 20% % Pss A 40 40 100 B 40 14 35 C 40 9 23 D 40 40 100 Investigtion Following the filed ISR test for metbolites B nd C, the smple nlysis ws put on hold nd n investigtion ws initited. Upon exmining the dt nd the study records, the following observtions nd conclusion were noted: Observtions Inference 1 The ISR dt for the prent compound A nd the metbolite D were in excellent greement with the originl dt (100% pssed criteri). The reproducibility for the metbolites B nd C ws concern. These suggested tht there ws no experimentl error from the originl runs nd the ISR runs. 2 For both metbolites B nd C, the ssy performnce ws better for qulity control (QC) smples thn for the study smples. The in-study precision nd ccurcy for metbolite B rnges from -11.1% to 1.4% nd 3.2% to 5.8% nd for metbolite C rnges from -1.5% to 3.0% nd 1.8% to 5.9%. These suggested tht there my be mtrix effect difference between the blnk mtrix plsm used to prepre the stndrds nd the QCs, nd the study smples mtrix. Overll, it indicted tht the mtrix effect presented in the study smples hs negtive impct to the ssy vribility for metbolites B nd C, but not for prent A nd metbolite D. One possible rtionle to ccount for mtrix effect difference cn be ttributed to the suitbility of the IS. One obvious solution to minimize the mtrix difference would be to revlidte the ssy with the SIL-IS for ll four nlytes. However, considering the time required for synthesis, revlidtion, nd the concern on the study smple stbility, n lterntive pproch is to dilute out the mtrix effect with the blnk mtrix used to prepre stndrd nd QCs with sufficiently lrge dilution fctor (e.g., 10). Although this pproch my be pplicble to metbolite B, it will not be pplicble to the minor metbolite C, s most smples fter 10-fold dilution will result in below quntifible limit (BQL) of metbolite C. With the sponsor s pprovl, the originl dt generted for metbolite C will only be used for informtion only. In order to test the bove hypothesis, some study smples were pooled to generte 20 individul smples. These pooled smples were ssyed twice for metbolite B using 10-fold dilution. The result of the dilution experiments showed tht the ssy vrition cused by mtrix in the study smple ws effectively minimized (Tble 3).

Tble 3. Dilution Experiments Showing the Mtrix Effect ws Minimlized for Metbolite B Anlyte B Investigtion Results # of INV smples % Pss 20 85.0 Smple ID 1st Anlysis Run ID 1st Anlysis Conc. (ng/ml) 2nd Anlysis Run ID 2nd Anlysis Conc. (ng/ml) % Difference Pooled Smple 1 14 574.841 15 528.847-8.3 Pooled Smple 2 14 800.218 15 677.711-16.6 Pooled Smple 3 14 314.190 15 275.811-13.0 Pooled Smple 4 14 499.461 15 434.986-13.8 Pooled Smple 5 14 2131.049 15 2006.878-6.0 Pooled Smple 6 14 2460.329 15 2298.377-6.8 Pooled Smple 7 14 3954.868 15 3446.033-13.8 Pooled Smple 8 14 2745.222 15 1892.830-36.8 Pooled Smple 9 14 2505.478 15 2970.975 17.0 Pooled Smple 10 14 3854.642 15 3059.831-23.0 Pooled Smple 11 14 4439.530 15 3842.291-14.4 Pooled Smple 12 14 2807.128 15 2536.163-10.1 Pooled Smple 13 14 2323.385 15 2129.957-8.7 Pooled Smple 14 14 2718.301 15 2743.594 0.9 Pooled Smple 15 14 1857.634 15 1787.170-3.9 Pooled Smple 16 14 12026.133 15 9809.399-20.3 Pooled Smple 17 14 8763.194 15 8032.055-8.7 Pooled Smple 18 14 7557.803 15 6918.018-8.8 Pooled Smple 19 14 15113.308 15 13029.597-14.8 Pooled Smple 20 14 6864.915 15 5865.826-15.7 % Difference = ((1 st Ressy 2 nd Ressy)/Averge of 1 st Ressy nd 2 nd Ressy) x 100 Subsequently, metbolite B ws ressyed from ll smples, except from the control groups, with 10-fold dilution. The relibility of the newly generted dt for metbolite B ws supported by successful ISR test (90.5% pss). Tble 4. ISR Summry for Metbolite B fter Investigtion Metbolite B ISR Results fter Investigtion # of ISR smples % Pss 42 90.5 Smple ID Originl Anlysis Run ID Originl Conc. (ng/ml) Repet Run ID Repet Conc. (ng/ml) % Difference PWL3F25 C Dy 7 1h 16 781.769 19 705.577-10.2 PWL3F25 C Dy 7 2h 16 2481.663 19 2196.643-12.2 PWL3F25 C Dy 7 6h 16 323.810 19 288.410-11.6 PWL3F25 C Dy 7 12h 16 993.882 19 964.899-3.0 PWL3F27 C Dy 7 2h 16 433.636 19 475.277 9.2 PWL3F27 C Dy 7 4h 16 716.982 19 840.253 15.8 PWL3F27 C Dy 7 6h 16 584.856 19 668.916 13.4 PWL3F27 C Dy 7 12h 16 621.049 19 603.029-2.9 PWLF328 C Dy 1 6h 16 455.283 19 443.403-2.6 PWL3F28 C Dy 7 2h 16 1366.715 19 1453.096 6.1 PWL4F32 D Dy 1 1h 18 4947.503 23 5334.462 7.5 PWL4F32 D Dy 1 2h 18 9936.796 23 10024.87 0.9 PWL4F32 D Dy 1 4h 18 9074.323 23 9806.04 7.8 PWL4F32 D Dy 1 6h 18 4529.285 23 4957.872 9.0 PWL4F32 D Dy 1 12h 18 18212.565 23 19175.217 5.1 PWL4F32 D Dy 1 24h 18 1225.992 23 1135.489-7.7 PWL4F32 D Dy 7 0h 18 1719.371 23 1635.151-5.0 PWL4F32 D Dy 7 1h 18 1585.534 23 1701.343 7.0 PWL4F32 D Dy 7 2h 18 3601.646 23 3553.559-1.3 PWL4F32 D Dy 7 4h 18 13289.471 23 14060.574 5.6 PWL4F32 D Dy 7 6h 18 11966.738 23 12011.100 0.4 PWL4F32 D Dy 7 12h 18 7332.635 23 7159.285-2.4 PWL4F32 D Dy 7 24h 18 2058.481 23 2921.901 34.7 PWL4M13 D Dy 1 6h 18 1904.360 23 2181.760 13.6 PWL4M13 D Dy 7 0h 18 2167.978 23 2234.993 3.0 PWL4M13 D Dy 7 1h 18 3629.342 23 4360.979 18.3 PWL4M13 D Dy 7 2h 18 3386.120 23 3791.866 11.3 PWL4M13 D Dy 7 24h 18 498.168 23 521.912 4.7 PWL4M14 D Dy 1 2h 18 527.782 23 504.219-4.6 PWL4M14 D Dy 1 4h 18 2378.377 23 2477.568 4.1 PWL4M14 D Dy 1 6h 18 2951.564 23 2855.399-3.3 PWL4M14 D Dy 1 24h 18 1072.793 23 1094.362 2.0 PWL4M14 D Dy 7 0h 18 532.724 23 459.522-14.8 PWL4M14 D Dy 7 12h 18 4264.303 23 4008.368-6.2 PWL4M14 D Dy 7 24h 18 1704.398 23 1410.889-18.8 PWL4M15 D Dy 7 0h 22 1066.822 23 1367.123 24.7 PWL4M15 D Dy 7 1h 22 1426.078 23 1234.661-14.4 PWL4M15 D Dy 7 2h 22 3238.436 23 2710.909-17.7 PWL4M15 D Dy 7 4h 22 5075.246 23 6289.010 21.4 PWL4M15 D Dy 7 6h 22 8287.302 23 5404.922-42.1 PWL4M15 D Dy 7 12h 22 3340.462 23 3037.136-9.5 PWL4M15 D Dy 7 24h 22 1336.250 23 1303.259-2.5 % Difference = ((1 st Ressy Originl)/Averge of Ressy nd Originl) x 100

>>> Incurred Smple Renlysis Findings >>> Mtrix effect presented in the study smples ws the root cuse of the ssy vribility for the metbolite B nd C uncovered by ISR test. >>> Hving pproprite IS, such s SIL-IS, is importnt to minimize the potentil mtrix effect presented in the study smples, which my not be present during vlidtion. >>> If the sitution llows, dilution cn be n effective pproch to minimize the mtrix effect presented in the study smples. >>> ISR ssessment provides dditionl nd importnt dt for the evlution of ssy performnce nd dt relibility. Cse Study 2 This cse study illustrtes the importnce of exmining the ISR results even if it meets the cceptnce criteri to help identify possible systemic errors. This is discreet ssy using stble isotope lbeled internl stndrd (SIL-IS) developed to support clinicl studies. The ccurcy nd precision for ssy during method vlidtion is -8.5% to 7.2% nd 1.3% to 15.0%, respectively. ISR Summry The ISR dt (Tble 5) met the cceptnce criteri. However, upon further review, it ws noted tht most of the dt (13 out of 14 smples) tht filed to meet the +/-20% criteri were from Run 1. Furthermore, 9 other smples in Run 1 were beyond +/-15%. In contrst, the dt from Runs 3 nd 11 mtched well with the corresponding dt from the ISR run (Run 12). These observtions brought the integrity of ll Run 1 dt into question.

>>> Incurred Smple Renlysis Tble 5. Compound X Incurred Smple Reproducibility Results 101,G2,D13,0h 1 2087.872 12 1856.708-11.7 101,G2,D14,0.5h 1 1610.092 12 1677.466 4.1 101,G2,D14,1.5h 1 1713.589 12 1629.247-5.0 101,G2,D14,2h 1 1652.765 12 1615.084-2.3 101,G2,D14,3h 1 2280.323 12 2289.742 0.4 101,G2,D14,4h 1 3014.757 12 3100.839 2.8 101,G2,D14,6h 1 3375.171 12 2836.305-17.4 101,G2,D14,8h 1 2641.475 12 2797.378 5.7 102,G2,D26,0h 1 979.537 12 753.719-26.1 102,G2,D27,0h 1 778.689 12 660.517-16.4 102,G2,D28,0h 1 751.627 12 647.900-14.8 102,G2,D28,0.5h 1 900.721 12 708.896-23.8 102,G2,D28,1h 1 916.106 12 810.244-12.3 102,G2,D28,1.5h 1 1184.087 12 940.851-22.9 102,G2,D28,2h 1 1365.693 12 1136.715-18.3 102,G2,D28,3h 1 1889.272 12 1620.436-15.3 102,G2,D28,4h 1 2029.748 12 1777.286-13.3 102,G2,D28,6h 1 1988.201 12 2028.915 2.0 103,G2,D14,0.5h 1 1085.773 12 906.951-17.9 103,G2,D14,1h 1 1103.389 12 881.442-22.4 103,G2,D14,1.5h 1 1248.380 12 1035.627-18.6 103,G2,D14,2h 1 1521.938 12 1262.261-18.7 103,G2,D14,3h 1 1709.981 12 1643.891-3.9 103,G2,D14,4h 1 2217.167 12 1735.754-24.4 103,G2,D14,6h 1 2715.513 12 2255.236-18.5 103,G2,D14,8h 1 1922.785 12 1628.443-16.6 108,G2,D27,0h 1 1438.522 12 1163.115-21.2 108,G2,D28,0h 1 1376.740 12 966.449-35.0 108,G2,D28,0.5h 1 1271.868 12 937.753-30.2 108,G2,D28,1h 1 1510.185 12 1095.369-31.8 108,G2,D28,1.5h 1 1991.747 12 1231.286-47.2 108,G2,D28,2h 1 2398.261 12 1579.958-41.1 108,G2,D28,3h 1 2656.306 12 2073.293-24.7 108,G2,D28,4h 1 2443.498 12 1806.356-30.0 110,G2,D14,0.5h 11 1954.422 12 2064.738 5.5 110,G2,D14,1h 11 2544.571 12 2891.901 12.8 110,G2,D14,1.5h 11 3594.114 12 3622.332 0.8 110,G2,D14,2h 11 3914.089 12 4316.082 9.8 110,G2,D14,3h 11 3922.751 12 4480.547 13.3 110,G2,D14,4h 11 3489.937 12 3657.187 4.7 110,G2,D14,6h 11 2815.752 12 3101.525 9.7 110,G2,D14,8h 11 2518.306 12 2899.777 14.1 111,G2,D28,0h 11 2025.106 12 2071.681 2.3 111,G2,D28,0.5h 11 1950.301 12 2229.870 13.4 111,G2,D28,1h 11 1848.487 12 2133.730 14.3 111,G2,D28,1.5h 11 1960.227 12 2057.687 4.9 111,G2,D28,2h 11 2292.712 12 2480.375 7.9 111,G2,D28,3h 11 3279.605 12 3483.927 6.0 111,G2,D28,4h 11 3140.516 12 3428.337 8.8 111,G2,D28,6h 11 2766.580 12 2902.813 4.8 112,G2,D14,0.5h 11 1774.783 12 1904.196 7.0 112,G2,D14,1h 11 2144.543 12 2183.799 1.8 112,G2,D14,1.5h 11 2329.472 12 2423.884 4.0 112,G2,D14,2h 11 2614.276 12 2728.521 4.3 112,G2,D14,3h 11 2637.009 12 2906.673 9.7 112,G2,D14,4h 11 2245.226 12 2540.706 12.3 112,G2,D14,6h 11 3859.618 12 4274.313 10.2 112,G2,D14,8h 11 3040.704 12 3311.100 8.5 117,G2,D28,3h 3 2877.225 12 2787.041-3.2 117,G2,D28,4h 3 2478.336 12 2746.166 10.3 117,G2,D28,6h 3 3107.050 12 3251.908 4.6 117,G2,D28,8h 3 3336.701 12 3297.377-1.2 117,G2,D28,10h 3 3524.980 12 3576.117 1.4 117,G2,D28,12h 3 3081.103 12 3110.655 1.0 117,G2,D28,16h 3 2241.527 12 2457.068 9.2 117,G2,29,24h 3 1940.758 12 1989.687 2.5 118,G2,D14,0.5h 3 1804.706 12 1797.490-0.4 118,G2,D14,1h 3 1790.069 12 2019.997 12.1 118,G2,D14,1.5h 3 2553.076 12 2887.415 12.3 118,G2,D14,2h 3 2595.955 12 2718.739 4.6 118,G2,D14,3h 3 3206.711 12 3080.784-4.0 118,G2,D14,4h 3 4645.108 12 3647.568-24.1 118,G2,D14,6h 3 3399.174 12 3439.400 1.2 118,G2,D14,8h 3 2720.020 12 2683.188-1.4 121,G2,D28,0h 3 2224.134 12 2257.100 1.5 121,G2,D28,0.5h 3 2209.616 12 2198.650-0.5 121,G2,D28,1h 3 2550.526 12 2858.880 11.4 121,G2,D28,1.5h 3 2704.991 12 2873.448 6.0 121,G2,D28,2h 3 3098.680 12 3334.080 7.3 121,G2,D28,3h 3 3752.064 12 3877.955 3.3 121,G2,D28,4h 3 4035.024 12 4383.122 8.3 121,G2,D28,6h 3 3688.969 12 3902.606 5.6 % Difference = ((Ressy Originl)/Averge of Ressy nd Originl) x 100 # of ISR smples % Pss 82 83.0 Smple ID Originl Anlysis Run ID Originl Conc. (ng/ml) Repet Run ID Repet Conc. (ng/ml) % Difference

>>> Incurred Smple Renlysis Investigtion To ddress the bove concern, n investigtion ws initited with the following points in mind: >>> All Run 1 dt were rejected. >>> All smples from Run 1 were ressyed in singlet. >>> The concentrtion dt from the ISR run (Run 12) were compred to the ressy run (Run 13) for ll 34 smples. >>> The pre-defined criteri re tht the ressy run will be considered vlid if more thn 2/3 of these 34 smples re within +/-20%. Tble 6 is the comprison of the ressyed dt nd the ISR dt, nd it clerly indicted tht the ressy run dt support the originl ISR dt. This lso indicted tht there ws n experimentl error from the originl Run (1). Tble 6. Compound X Incurred Smple Reproducibility Results Smple ID ISR Anlysis Run ID Originl Conc. (ng/ml) Repet Run ID Repet Conc. (ng/ml) % Difference 101,G2,D13,0h 12 1856.708 13 1833.034-1.3 101,G2,D14,0.5h 12 1677.466 13 1496.129-11.4 101,G2,D14,1.5h 12 1629.247 13 1628.728 0.0 101,G2,D14,2h 12 1615.084 13 1657.001 2.6 101,G2,D14,3h 12 2289.742 13 2329.425 1.7 101,G2,D14,4h 12 3100.839 13 2982.349-3.9 101,G2,D14,6h 12 2836.305 13 3090.292 8.6 101,G2,D14,8h 12 2797.378 13 2613.182-6.8 102,G2,D26,0h 12 753.719 13 793.158 5.1 102,G2,D27,0h 12 660.517 13 705.364 6.6 102,G2,D28,0h 12 647.900 13 717.052 10.1 102,G2,D28,0.5h 12 708.896 13 697.535-1.6 102,G2,D28,1h 12 810.244 13 833.202 2.8 102,G2,D28,1.5h 12 940.851 13 960.815 2.1 102,G2,D28,2h 12 1136.715 13 1187.352 4.4 102,G2,D28,3h 12 1620.436 13 1722.059 6.1 102,G2,D28,4h 12 1777.286 13 1752.482-1.4 102,G2,D28,6h 12 2028.915 13 2022.997-0.3 103,G2,D14,0.5h 12 906.951 13 898.783-0.9 103,G2,D14,1h 12 881.442 13 947.033 7.2 103,G2,D14,1.5h 12 1035.627 13 998.804-3.6 103,G2,D14,2h 12 1262.261 13 1226.231-2.9 103,G2,D14,3h 12 1643.891 13 1439.689-13.2 103,G2,D14,4h 12 1735.754 13 1688.747-2.7 103,G2,D14,6h 12 2255.236 13 2224.675-1.4 103,G2,D14,8h 12 1628.443 13 1569.432-3.7 108,G2,D27,0h 12 1163.115 13 1142.753-1.8 108,G2,D28,0h 12 966.449 13 974.496 0.8 108,G2,D28,0.5h 12 937.753 13 991.519 5.6 108,G2,D28,1h 12 1095.369 13 1146.245 4.5 108,G2,D28,1.5h 12 1231.286 13 1229.703-0.1 108,G2,D28,2h 12 1579.958 13 1522.282-3.7 108,G2,D28,3h 12 2073.293 13 1897.349-8.9 108,G2,D28,4h 12 1806.356 13 1742.164-3.6 % Difference = ((Ressy Originl)/Averge of Ressy nd Originl) x 100 % Pss 100.0 >>> >>> >>> Findings A successful ISR ssessment helps improve confidence on the performnce of the ssy nd relibility of dt, but it does not men tht the work required to ensure dt qulity is complete. Crefully exmining the ISR results, even if it meets the cceptnce criteri, my help identify possible systemic errors. It is importnt to look beyond the cceptnce criteri from n ISR to further improve the confidence of the dt. Delwre Technology Prk 3 Innovtion Wy, Suite 240 Newrk, DE 19711 web www.qps-us.com TEL (302) 369-5601 Fx (302) 369-5602