SecondLook Digital for Hologic Labeling and User Manual
Contents
1. 10 ADVERSE EEFEGTS 10 CLINICAL STUDIES ciet eene de Ea rie rese et to eerie ed eius eor ee bie Evita ced 10 DETAILED DEVICE DESCRIPTION esses enne nnne nnn tnnt trennen nnn tren nnn nnns 17 CONFORMANCE TO STANDARDS 21 HOW SUPPLIED EET 21 22 RADIOLOGIST REVIEW PRIOR TO VIEWING CAD MARKS sese 22 RADIOLOGIST REVIEW WITH CAD MARKS c2 csccceceeseeeeceeeeseeeeceaecaeeeeeeaecaaesaesaeseaeeaeseseaeeaeeneeeaes 22 23 TRAINING INSTRUCTIONS P 23 SAMPUE 24 E E E 27 28 1717 3009 Rev G iCAD Inc Page 5 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G This page intentionally left blank 1717 3009 Rev G iCAD Inc Page 6 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G This manual describes the SecondLook Digital Computer Aided Detection CAD system and provides training to radiologists using the SecondLook Digital system for breast cancer detection e Section 2 provides SecondLook device labeling e Section 3 describes how a radiologist should use SecondLook Digital e Section 4 provides a sample case to familiarize the radiologist with SecondLook Digital e Section 5 provides a summar2. SecondLook Digital for Hologic Labeling and User Manual Rev G Did not compute 95 Cls for per image FP amp Mark Rates Every non cancer case had 4 standard images so case based FP amp Mark Rates are more clinically relevant Table 4 and per image 95 Cls are not needed 2 7 Detailed Device Description SecondLook uses computer aided detection CAD algorithms to identify regions of interest on mammograms that may contain suspicious finding The CAD algorithms use advanced image processing feature computations and pattern recognition technology to analyze the images for potential areas of concern These potential areas of concern are displayed for the radiologist by overlaying CAD marks at the appropriate locations of the mammography images within the softcopy review workstation or on a paper printout The CAD marks are used by the radiologist as an additional tool in breast cancer detection An overview of the SecondLook CAD algorithms is shown in Figure 1 Standard Mammography Images MicroCalc Density Algorithm Algorithm Calc Image Density Image Enhancement Enhancement MicroCalc Density Detector Detector 1 Region Clustering Growing MicroCalc Density Classifier Classifier L L Context Based Patient Evaluation Areas of Concern Highlighted by CAD Marks Figure 1 SecondLook CAD Algorithms Overview
3. 86 79 92 Table 3 Image Based Sensitivity Results for SecondLook Digital on Hologic Digital Images Selenia Selenia Dimensions Total Images Enrolled 200 198 Images with Clearly Identifiable Lesion s 200 198 Predominant Type of Most Suspicious Lesion Images with Masses 148 115 Images with Microcalcifications 21 79 Images with Mixed Masses Microcalcifications 31 4 Scoring Results True Positive Images With Masses Percentage Detected Medium 95 Confidence Interval 73 65 80 64 56 73 Percentage Detected High 95 Confidence Interval 73 66 80 69 60 77 With Microcalcifications Percentage Detected Medium 95 Confidence Interval 100 100 100 71 61 81 Percentage Detected High 95 Confidence Interval 100 100 100 71 61 81 With Mixed Masses Microcalcifications Percentage Detected Medium 95 Confidence Interval 94 85 100 100 25 100 Percentage Detected High 95 Confidence Interval 97 90 100 100 25 100 Overall 1717 3009 Rev G iCAD Inc Page 14 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G Percentage Detected Medium 95 78 73 84 68 61 74 Confidence Interval Percentage Detected High 95 Confidence 80 74 85 70 64 76 Inte
4. Percentage Number of Mixed Mass Microcalcification Cases 12 15 2 2 Percentage Average Pathology Mass Lesion size cm a Range Minimum Maximum 0 6 4 8 0 3 3 5 1717 3009 iCAD Inc Page 11 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G SecondLook Standalone Testing with Hologic Images Standalone testing of the SecondLook version 7 2 with Hologic images provides a performance measure i e sensitivity and average number of false positives per image or case in the absence of any interaction with a radiologist Standalone performance measures how often the CAD device places prompts over regions that contain or do not contain known breast abnormalities i e microcalcifications and or masses in the absence of radiologist interaction Sensitivity Analysis The sensitivity of SecondLook version 7 2 with Hologic images was estimated using electronic truth and scoring that was manually confirmed by a radiologist Electronic truth consists of an experienced mammography radiologist drawing an electronic truth box on each lesion using pertinent data from the mammography and pathology reports The lesion type mass microcalcification mixed is also noted Electronic scoring takes the CAD marks generated by SecondLook and compares their locations to the electronic truth boxes to determine if each CAD mark is a cancer hit or a false positive A
5. SecondLook Digital for Hologic Labeling and User Manual Rev G Table 2 Case Based Sensitivity Results for SecondLook Digital on Hologic Digital Images Selenia Selenia Dimensions Total Cases Enrolled 80 100 Cases with Clearly Identifiable Lesion s 80 100 Case Type Cases with Masses 59 59 Cases with Microcalcifications 9 39 Cases with Mixed Masses Microcalcifications 12 2 Scoring Results True Positive Cases With Masses Percentage Detected Medium Confidence Interval 95 90 82 98 83 73 92 Percentage Detected High 95 Confidence Interval 90 82 98 86 76 95 With Microcalcifications Percentage Detected Medium Confidence Interval 95 100 100 100 85 72 95 Percentage Detected High 95 Confidence Interval 100 100 100 85 72 95 With Mixed Masses Microcalcifications Percentage Detected Medium Confidence Interval 95 92 76 100 100 0 100 Percentage Detected High 95 Confidence Interval 92 75 100 100 0 100 Overall Percentage Detected Medium 95 91 85 98 84 76 91 1717 3009 Rev G iCAD Inc Page 13 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G Confidence Interval Percentage Detected High 95 Confidence Interval 91 85 98
6. Digital CAD marks with the softcopy review workstation after determining whether or not a work up is indicated from her or his initial review of the patient mammograms The radiologist will take a SecondLook at the mammograms corresponding to any CAD marks From this re evaluation of the mammograms the radiologist determines if any additional work up is required If there are no CAD marks no re evaluation of the mammograms is necessary Work up decisions are not based solely upon the CAD marks All work up decisions are based upon review of the mammograms supporting clinical information and CAD marks by the radiologist Areas of concern marked by SecondLook Digital include suspicious clusters of microcalcifications spiculated and non spiculated masses architectural distortions and focal asymmetric densities Below is the recommended case review process with SecondLook Digital 1 Review patient history and evaluate Hologic mammograms prior to turning on and viewing CAD marks with softcopy review workstation 2 Make initial interpretation 3 Turn on and view CAD marks with softcopy review workstation and identify potential areas of concern 4 Review mammograms re evaluating areas of concern highlighted by CAD marks with softcopy review workstation 5 Render decision It is very important to remember that it is the radiologist who makes the final decision about a case When a radiologist decides to work up a case the CAD mar
7. iCADs Never stop looking SecondLook Digital for Hologic Labeling and User Manual 2014 iCAD Inc All Rights Reserved iCAD the iCAD logo Never Stop Looking and SecondLook are registered trademarks of iCAD Inc Other company product and service names may be trademarks or service marks of others 1717 3009 Rev G SecondLook Digital for Hologic Labeling and User Manual Rev G This page intentionally left blank 1717 3009 Rev G iCAD Inc Page 2 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G Never stop looking 98 Spit Brook Rd Suite 100 Nashua NH 03062 USA 603 882 5200 The European Representative for iCAD Inc is MDSS GmbH Schiffgraben 41 30175 Hannover Germany 1717 3009 Rev G iCAD Inc Page 3 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G This page intentionally left blank 1717 3009 Rev G iCAD Inc Page 4 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G 2 1 2 2 2 3 2 4 2 5 2 6 2 7 2 8 2 9 3 1 3 2 4 1 4 2 7 7 INDICATIONS FOR USE ie access eret eee he Facets Rede Hen ede gU Pr ep OR E dde 7 BRIEF DEVICE re ee a bn teu 7 MPMcId 8 meus
8. system any resulting CAD marks should not be used by the radiologist in evaluating the patient The performance of the SecondLook V7 2 device has not been established for views ML views and breast implants and may differ from those derived using conventional views of the breast i e CC and MLO views Precautions Installation and Maintenance e This product contains no independently user serviceable parts To prevent damage to the system do not attempt to install or repair the SecondLook system Only trained personnel are qualified to install or repair the system For service training contact iCAD Inc at 1 866 280 2239 e Disconnect power cord before moving or servicing 2 5 Adverse Effects SecondLook may increase your false positive rates for both screening and diagnostic mammography Increased false positives may lead to unnecessary additional imaging radiation exposure biopsy patient anxiety etc 2 6 Clinical Studies Refer to the SecondLook Analog for further details regarding the testing studies used to support the safety and effectiveness of the original approval of the SecondLook analog device for use with digitized film screen mammograms 1717 3009 Rev G iCAD Inc Page 10 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G Benchmark testing Benchmark testing consisted of a limited standalone analysis i e analysis of the device without radiologist interaction on sample of Holo
9. of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G 16 Dean JC Ilvento CC Improved Cancer Detection Using Computer Aided Detection With Diagnostic and Screening Mammography Prospective Study of 104 Cancers AJR 2006 187 20 28 17 Ko JM Nicholas MJ Mendel JB Slanetz PJ Prospective Assessment of Computer Aided Detection in Interpretation of Screening Mammography AJR 2006 187 1483 1491 18 Fenton JJ Taplin SH Carney PA et al Influence of Computer Aided Detection on Performance of Screening Mammography NEJM 2007 356 1399 1409 19 Georgian Smith D Moore RH Halpern E et al Blinded Comparison of Computer Aided Detection With Human Second Reading in Screening Mammography AJR 2007 189 1135 1141 20 Gromet M Comparison of Computer Aided Detection to Double Reading of Screening Mammograms Review of 231 221 Mammograms AJR 2008 190 854 859 21 Brem RF Baum J Lechner M et al Improvement in Sensitivity of Screening Mammography With Computer Aided Detection A Multiinstitutional Trial 2003 181 687 693 22 Destounis SV DiNitto P Logan Young W et al Can Computer Aided Detection With Double Reading of Screening Mammograms Help Decrease the False Negative Rate Initial Experience Radiology 2004 232 578 584 23 Gilbert FJ Astley SM McGee MA et al Single Reading with Computer Aided Detection and Double Reading of Screening Mammograms in the United Kingdom National Breast Screen
10. 1717 3009 Rev G iCAD Inc Page 17 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G The CAD algorithms begin with image enhancement of the digitized mammographic images to accentuate all areas that could be individual microcalcifications and densities In the case of directly acquired images the digital images are first transformed into images that resemble digitized film in order to accommodate variations in inter pixel spacing gray level mapping and bit depth It should be noted that the Modulation Transfer Function MTF for Hologic images deviates from the MTF specified for SecondLook in the high frequency range While the MTF is not directly used in the calculations performed by SecondLook this deviation may impact the calculation of subtle features along the margins of lesions The microcalcification and density detectors then identify the areas that are most likely to be individual microcalcifications and densities based on an initial analysis of morphological and intensity measurements The types of densities detected are depicted in Figure 2 and include spiculated and non spiculated masses architectural distortions and focal densities 1717 3009 Rev G iCAD Inc Page 18 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G Circumscribed Masses Round Oval Lobular Irregular Mass with Microlobulated Mass Obscured Mass Indistinct Margins CON NO Spiculated Mass Architectural Di
11. CAD mark is assessed as a true positive detection if the CAD mark of a given type mass microcalcification hits a truth box of the same type mixed lesions can be hit by a CAD mark of either type A microcalcification CAD mark is a hit if the CAD mark has any overlap with a microcalcification or mixed truth box A mass CAD mark is a hit if its center falls within a mass or mixed truth box Any CAD mark that is not a hit is counted as a false positive Sensitivity is a count of the true positive detection divided by the total number of cases with mammographically visible cancer and was calculated separately on a per case i e per patient and per images i e mammographic view basis True positive detection was defined as follows For each scoring method if any lesion on an image was scored as a hit answered Yes then the image was scored as a true positive and the case was scored as a true positive Per case sensitivity was computed as the number of true positive cases divided by the total number of cases with mammographically visible cancer with 95 confidence intervals Per image sensitivity was computed as the number of true positive images divided by the total number of images with mammographically visible cancer with 95 The 95 Cl were computed using a statistical resampling technique 5000 sample bootstrapping Results of the sensitivity analysis are summarized in Tables 2 and 3 1717 3009 Rev G iCAD Inc Page 12 of 29
12. Look the radiologist must use their interpretative skills to determine if the area should be worked up based on its mammographic appearance e SecondLook is neither designed nor intended to prompt to o interval change s between mammographic exams asymmetry between the left and right breast tubular density solitary dilated duct skin thickening or nipple retraction O O O O 1717 3009 Rev G iCAD Inc Page 8 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G Warnings System Operation A e Do not use the SecondLook system if you suspect any electrical component is defective or inoperable Do not place liquids on or near SecondLook If a liquid is accidentally spilled on electrical components immediately turn off the computer which will automatically shut down the system to prevent any potential electrical shock Contact iCAD Inc at 1 866 280 2239 for further instructions e Ensure that the system is connected to a properly wired and grounded power receptacle e Ensure that the voltage and current requirements are within system specifications to avoid bodily injury from electrical shock or fire hazard Warnings Installation and Maintenance A e EMC Warning This SecondLook system has been tested and found to comply with IEC 60950 1 EN 55022 and EN 55024 This system generates uses and can radiate radio frequency energy and if not installed and used in accordance with our installation instr
13. Y REVIEW WORKSTATION FROM THIS RE EVALUATION OF THE MAMMOGRAMS THE RADIOLOGIST MAKES THE FINAL WORK UP ica DECISION UN 1717 3009 Rev G iCAD Inc Page 25 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G Case Summary History 62 yo female with palpable mass in upper outer quadrant of right breast No family history of breast cancer Mammographic findings 3 cm circumscribed mass with partially obscured borders in the right breast at 10 o clock shown to be a cyst on ultrasound Linear distribution of pleomorphic calcifications in the right breast at 2 o clock posteriorly Pathology Ductal carcinoma in situ arrows show location 1717 3009 Rev G iCAD Inc Page 26 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G The radiologist uses the SecondLook Digital CAD marks with mammography according to the following steps 1 The radiologist first reviews the Hologic mammograms without viewing the CAD marks following her or his existing procedures of clinical practice The radiologist will make an initial determination if a work up is indicated for the patient prior to turning on and viewing the CAD marks with the softcopy review workstation 2 The radiologist turns on and views the CAD marks with the softcopy review workstation after determining whether or not a work up is indicated from her or his initial review of the patient mammograms 3 The radiologi
14. e the CAD SecondLook with the Hologic FFDM system is intended to be used by a radiologist as follows The radiologist must always first perform a full conventional read of the mammogram and only after completing the conventional read the radiologist may choose to display the CAD marks which may prompt to areas that were or were not examined during the first read It is crucial to understand that 99 6 of all CAD marks will be placed over areas that are normal breast tissue or benign findings Be aware that the SecondLook is not a diagnostic device as the CAD marks are intended to be used to assist only in detection and not to assist with interpretation 2 3 Warnings A Warnings Radiological Interpretation e The radiologist must always first perform a full conventional read of the mammogram and only after completing the conventional read the radiologist may choose to display the CAD marks which may prompt to areas that were or were not examined during the first read e The presence or absence of a CAD mark should not in any manner influence your diagnostic decision as to the nature of a mammographic finding i e normal vs benign vs malignant or the clinical action to be taken e g additional imaging or biopsy e Donotrely on the size or shape of the CAD mark as it may not be representative of the actual extent or shape of the breast lesion Upon re evaluation for the original mammogram at the locations indicated by Second
15. gic mammograms e Selenia 80 cancers with 59 masses 9 microcalcification clusters and 12 mixed mass microcalcifications e Selenia Dimensions 100 cancers with 59 masses 39 microcalcification clusters and 2 mixed mass microcalcifications Note that standalone performance testing of SecondLook version 7 2 on Hologic images cannot be directly compared to standalone performance testing of SecondLook on digitized film screen images The benchmark testing did not measure the effect of the device on radiologist performance and cannot measure or predict any change in radiologist s cancer detection rates when using the device as intended Hologic Database Description Table 1 provides the mammographic and pathologic characteristics on the Hologic cases The database of Selenia images included 80 cases with cancer and 95 free of cancer Of the cancer cases there were 59 mass cases 22 2 2 0 cm 9 microcalcification cases 1 2 2 0 cm and 12 mixed mass microcalcification cases 8 2 2 0cm The database of Selenia Dimensions images included 100 cases with cancer and 140 free of cancer Of the cancer cases there were 59 mass cases 15 2 2 0 cm 39 microcalcification cases 10 2 2 0 cm and 2 mixed mass microcalcification cases 1 2 2 0cm Table 1 Hologic Database Demographics Cancers Selenia Selenia Dimensions Number of Primarily Mass Cases Percentage 59 74 59 59 Number of Primarily Microcalcification Cases 9 11 39 39
16. ing Program Radiology 2006 241 47 53 24 Balleyguier C Kinkel K Fermanian J et al Computer Aided Detection CAD in Mammography Does it Help the Junior or the Senior Radiologist European Journal of Radiology 2005 54 90 96 25 Marx C Malich A Facius M et al Are Unnecessary Follow up Procedures Induced by Computer Aided Diagnosis CAD in Mammography Comparison of Mammographic Diagnosis With and Without Use of CAD European Journal of Radiology 2004 51 66 72 26 Hukkinen K Vehmas T Pamilo M Kivisaari L Effect of Computer Aided Detection on Mammographic Performance Experimental Study onReaders With Different Levels of Experience Acta Radiologica 2006 47 257 263 27 Taplin SH Rutter CM Lehman CD Testing the Effect of Computer Assisted Detection on Interpretive Performance in Screening Mammography AJR 2006 187 1475 1482 28 Brem RF Baum J Kaplan S Improvement in Sensitivity of Screening Mammography with Computer Aided Detection A Multiinstitutional Trial AJR 2003 181 687 693 1717 3009 Rev G iCAD Inc Page 29 of 29
17. ks must not change the decision however the CAD marks can identify locations for further work up that were initially undetected by the radiologist 1717 3009 Rev G iCAD Inc Page 22 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G 41 Training Instructions One sample case demonstrates the use of SecondLook Digital for the radiologist prior to clinical use This case is intended to familiarize the radiologist with the procedures for using the SecondLook Digital CAD marks The case review procedures are emphasized Therefore the training is accomplished by following the case presentation in Section 4 2 of this manual without requiring use of the softcopy review station For the example case in the manual the procedures for using SecondLook Digital CAD marks are demonstrated to the radiologist with the following steps 1 The first page will provide the case history and printed versions of the Hologic mammograms without CAD marks During clinical use the radiologist would first review the mammograms without viewing the CAD marks following her or his existing procedures of clinical practice The radiologist would make an initial determination if a work up were indicated for the patient prior to turning on and viewing the CAD marks with the softcopy review workstation 2 The second page contains printed versions of the mammograms with CAD marks turned on During clinical use the radiologist would take a SecondLook a
18. rval Marker Rate Analysis The false marker rate analysis of the SecondLook version 7 2 with Hologic images was measured in two separate ways 1 Averaging the number of marks in all 4 views of a set of BI RADS 1 cases and 2 By averaging the number of marks in all 4 views of a separate set of screening cases without proven cancers all BI RADS For this analysis all Hologic images were assumed not to contain a mammographically visible cancer and therefore all CAD marks were considered false positive marks This includes marks on benign or otherwise mammographically interesting regions that are ultimately determined to not be cancer The false marker rate was calculated per case i e per patient and per image i e mammographic view separately Associated 95 Cl were estimated using a statistical resampling technique 5000 sample bootstrapping Per case false marker rate calculation is simply the total number of false positive marks for the case Per image false marker rate calculation is the total number of false positive marks divided by the total number of mammographic views The results for the false marker rate are summarized in Tables 4 and 5 Table 4 Case Based False Marker Rate Calculations for SecondLook Digital on Hologic Digital Images Selenia Selenia Dimensions Cases Enrolled Total 95 140 Contained Standard 4 Views Screening 55 140 Dataset Contained S
19. sed by microcalcifications and density classifiers which are specifically designed to select the areas most likely to have features that may be seen with cancer Further analysis uses the context of all areas selected for the patient For example there is a maximum total number of SecondLook CAD marks each 4 image case can include Simultaneous analysis of all areas of concern detected in the patient allows the locations most likely to be cancer to be highlighted by the CAD marks 1717 3009 Rev G iCAD Inc Page 20 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G Figure 4 Region Growing to Determine Shape of Density 2 8 Conformance to Standards Refer to the SecondLook Digital Service Manual for the CE Declaration of Conformity DTBO60 2 9 How Supplied The SecondLook system includes the following components e Computer 1717 3009 Rev G iCAD Inc Page 21 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G 3 1 3 2 Radiologist Review Prior to Viewing CAD Marks The radiologist first reviews the Hologic mammograms without viewing the SecondLook Digital CAD marks following her or his existing procedures of clinical practice The radiologist will make an initial determination if a work up is indicated for the patient prior to turning on and viewing the CAD marks with the softcopy review workstation Radiologist Review with CAD Marks The radiologist turns on and views the SecondLook
20. st will take a SecondLook at the mammograms corresponding to any CAD marks From this re evaluation of the mammograms the radiologist determines if any additional work up is required If there are no CAD marks no re evaluation of the mammograms is necessary Work up decisions are not based solely upon the CAD marks All work up decisions are based upon review of the mammograms supporting clinical information and CAD marks by the radiologist 1717 3009 Rev G iCAD Inc Page 27 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G 1 Bird RE Wallace TW Yankaskas BC Analysis of Cancers Missed at Screening Mammography Radiology 1992 184 613 617 2 Sickles EA Auditing Your Practice RSNA Categorical Course in Breast Imaging 1995 pp 81 91 3 Harvey JA Fajardo LL Innis CA Previous Mammograms in Patients with Impalpable Breast Carcinoma Retrospective vs Blinded Interpretation AJR 1993 161 1167 1172 4 Martin JE Moskowitz M Milbrath JR Breast Cancer Missed by Mammography AJR 1979 132 737 739 5 Schmidt RA Nishikawa RM Digital Screening Mammography PPO Updates 8 7 1 16 1994 6 Thurfjell EL Lernevall KA Taube AAS Benefit of Independent Double Reading Population based Mammography Screening Program Radiology 1994 191 241 244 7 Economic Impact Analysis of Regulations Under the Mammography Quality Standards Act of 1992 U S Food and Drug Administration and Eastern Research Gro
21. stortion JA Figure 2 Densities Detected by SecondLook Further analysis of detected areas is accomplished by clustering individual microcalcifications and region growing densities Clusters include 3 or more individual microcalcifications that are each no more than 4 1 millimeters apart Figure 3 depicts portions of three different mammography images showing how the SecondLook system would highlight microcalcifications clusters in these examples These examples use CAD marks that are rectangular and correspond to the approximate size of the microcalcifications Region growing determines the shape of potential densities as shown in Figure 4 1717 3009 Rev G iCAD Inc Page 19 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G Figure 3 CalcMarks Highlighting Microcalcifications Clusters with a The minimum number of calcifications 0 The extent of the CalcMark enclosing all calcifications considered as part of the cluster c Overlapping CalcMarks are distinctly highlighted even when clusters are close to each other After clustering for microcalcifications analysis and region growing for density analysis clinically relevant and mathematical features are then computed to describe each detected cluster of microcalcifications and density For example the variability in size and shape of the calcifications in a cluster are good features to describe clusters of microcalcifications These features are u
22. t the mammograms corresponding to any CAD marks From this re evaluation of the mammograms the radiologist would determine if any additional work up was required If there were no CAD marks no re evaluation of the mammograms would be necessary Work up decisions are not based solely upon the CAD marks All work up decisions are based upon review of the mammograms supporting clinical information and CAD marks by the radiologist 3 The third page then presents a summary of the case which includes the case history the mammographic findings and the resulting pathology An arrow points to the location of the tumor in printed versions of the mammograms 1717 3009 Rev G iCAD Inc Page 23 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G 4 2 Sample Case Case History and Mammograms History 62 yo female with palpable mass in upper outer quadrant of right breast No family history of breast cancer DURING CLINICAL USE THE INITIAL MAMMOGRAPHY REVIEW AND INITIAL WORK UP DECISION WOULD ACCOMPLISHED oe 1717 3009 Rev G iCAD Inc Page 24 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G Mammograms with CAD Marks Note The softcopy review workstation may use symbols other than rectangles calcifications and ellipses masses for the CAD marks DURING CLINICAL USE THE AREAS OF CONCERN HIGHLIGHTED BY THE CAD MARKS WOULD BE RE EVALUATED USING THE SOFTCOP
23. tandard 4 Views BIRADS 1 40 100 Dataset 1717 3009 Rev G iCAD Inc Page 15 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G Scoring Results Screening Dataset Confidence Interval Average CAD Marks per Case Medium 2 25 1 70 2 82 2 3 1 9 2 6 95 Confidence Interval Average CAD Marks per Case High 95 2 49 1 90 3 07 2 8 2 4 3 1 Confidence Interval BIRADS 1 Dataset Average CAD Marks per Case Medium 2 10 1 47 2 73 2 0 1 6 2 3 95 Confidence Interval Average CAD Marks per Case High 95 2 38 1 72 3 03 2 5 2 1 2 8 Table 5 Image Based False Marker Rate Calculations for SecondLook Digital on Hologic Digital Images 95 Confidence Interval Selenia Selenia Dimensions Images Enrolled Total 380 560 Contained Standard 4 Views Screening 220 560 Dataset Contained Standard 4 Views BIRADS 1 160 400 Dataset Scoring Results Screening Dataset Average CAD Marks per Image Medium 0 56 0 46 0 67 0 6 9596 Confidence Interval Average CAD Marks per Image High 0 62 0 51 0 73 0 7 95 Confidence Interval BIRADS 1 Dataset Average CAD Marks per Image Medium 0 52 0 40 0 65 0 5 95 Confidence Interval Average CAD Marks per Image High 0 59 0 46 0 72 0 6 1717 3009 Rev G iCAD Inc Page 16 of 29
24. uctions may cause or be subject to harmful interference with other devices in the vicinity If the SecondLook system appears to cause or be subject to harmful interference try the following steps to correct the problem o Reorient or relocate the SecondLook system or the interface device o Increase the separation between the SecondLook system and the interfering device Plug the SecondLook system into an outlet on a different circuit from the interfering device o Contact iCAD Inc at 1 866 280 2239 for further instructions e Temperature and Humidity Warning SecondLook system operations must be performed within the following temperature and humidity ranges o Temperature 50 95 Fahrenheit 10 35 Celsius o Humidity 20 80 1717 3009 Rev G iCAD Inc Page 9 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G 2 4 Precautions A Precautions System Operation e prevent damage to the system maintain equipment in a well ventilated air conditioned environment e Effectiveness and safety in patients with breast implants has not been established for views that include the implant When non displaced implant views are analyzed by the system any resulting CAD marks should not be used by the radiologist in evaluating the patient e Effectiveness and safety have not been established for non standard mammographic views e g magnification compression views When these views are analyzed by the
25. up Inc Task Order No 1 Contract No 223 94 8031 October 7 1997 8 Quality Determinants of Mammography Clinical Practice Guideline Number 13 Agency for Health Care Policy and Research Publication No 95 0632 October 1994 9 Warren Burhenne LJ Wood SA D Orsi CJ et al Potential Contribution of Computer Aided Detection to the Sensitivity of Screening Mammography Radiology 2000 215 554 562 10 Freer TW Ulissey MJ Screening Mammography with Computer Aided Detection Prospective Study of 12 860 Patients in a Community Breast Center Radiology 2001 220 781 786 11 Gur D Sumkin JH Rockette HE et al Changes in Breast Cancer Detection and Mammography Recall Rates After the Introduction of a Computer Aided Detection System JNCI 2004 96 3 185 190 12 Birdwell RL Bandodkar P Ikeda DM Computer Aided Detection With Screening Mammography in a University Hospital Setting Radiology 2005 236 451 457 13 Cupples TE Cunningham JE Reynolds JC Impact of Computer Aided Detection in a Regional Screening Mammography Program AJR 2005 185 944 950 14 Khoo LAL Taylor P Given Wilson RM Computer Aided Detection in the United Kingdom National Breast Screening Programme Prospective Study Radiology 2005 237 444 449 15 Morton MJ Whaley DH Brandt KR et al Screening Mammograms linterpretation With Computer Aided Detection Prospective Evaluation Radiology 2006 239 375 383 1717 3009 Rev G iCAD Inc Page 28
26. y of the radiologist use of SecondLook Digital e Section 6 provides a list of clinical references 2 1 Indications for Use The SecondLook Computer Aided Detection CAD system for mammography is intended to identify and mark regions of interest on screening and diagnostic mammograms from Hologic Selenia and Selenia Dimensions full field digital mammography FFDM systems to bring them to the attention of the radiologist after an initial reading has been completed Thus the system prompts the radiologist to areas on Hologic mammograms for second review only 2 2 Brief Device Description SecondLook is a mammographic CAD system that prompts radiologists to areas on Hologic mammograms for a second review only The CAD algorithm version 7 2 includes image processing feature computations and pattern recognition technology to detect regions of interest The algorithm was originally trained on digitized film screen mammograms and intended to more specifically identify potential breast lesions appearing as clusters of microcalcifications and or masses The CAD system was adapted to run on Hologic images but the CAD algorithm design remained unchanged and was not otherwise retrained on the Hologic mammograms 1717 3009 Rev G iCAD Inc Page 7 of 29 SecondLook Digital for Hologic Labeling and User Manual Rev G For hardcopy reading the SecondLook output can be presented on a paper printout showing the CAD marks within the mammogram How to Us
Download Pdf Manuals
Related Search