Publications

Matching entries: 0
settings...
Samulski, M. & Karssemeijer, N. (2011), “Optimizing Case-based Detection Performance in a Multiview CAD System for Mammography”, IEEE Transactions on Medical Imaging, Apr, 2011. Vol. 30(4), pp. 1001-1009.

Abstract: When reading mammograms, radiologists combine information from multiple views to detect abnormalities. Most computer-aided detection (CAD) systems, however, use primitive methods for inclusion of multi-view context or analyze each view independently. In previous research it was found that in mammography lesion-based detection performance of CAD systems can be improved when correspondences between MLO and CC views are taken into account. However, detection at case level detection did not improve. In this paper, we propose a new learning method for multi-view CAD systems, which is aimed at optimizing case-based detection performance. The method builds on a single-view lesion detection system and a correspondence classifier. The latter provides class probabilities for the various types of region pairs and correspondence features. The correspondence classifier output is used to bias the selection of training patterns for a multi-view CAD system. In this way training can be forced to focus on optimization of case-based detection performance. The method is applied to the problem of detecting malignant masses and architectural distortions. Experiments involve 454 mammograms consisting of 4 views with a malignant region visible in at least one of the views. To evaluate performance, 5-fold cross validation and FROC analysis was performed. Bootstrapping was used for statistical analysis. A significant increase of case-based detection performance was found when the proposed method was used. Mean sensitivity increased by 4.7% in the range of 0.01-0.5 false positives per image.
BibTeX:

  @article{Samu11a,
    author = {Samulski, Maurice and Karssemeijer, Nico},
    title = {Optimizing Case-based Detection Performance in a Multiview CAD System for Mammography},
    journal = {IEEE Transactions on Medical Imaging},
    year = {2011},
    volume = {30},
    number = {4},
    pages = {1001--1009},
    url = {http://dx.doi.org/10.1109/TMI.2011.2105886},
    doi = {http://dx.doi.org/10.1109/TMI.2011.2105886}
  }
  
Samulski, M., Snoeren, P., Platel, B., van Ginneken, B., Hogeweg, L., Schaefer-Prokop, C. & Karssemeijer, N. (2011), “Computer-Aided Detection as a Decision Assistant in Chest Radiography”, In Medical Imaging, Proceedings of the SPIE. Volume 7966, pp. 796614.

Abstract: Background: Contrary to what may be expected, finding abnormalities in complex images like pulmonary nodules in chest radiographs is not dominated by time-consuming search strategies but by an almost immediate global interpretation. This was already known in the nineteen-seventies from experiments with briefly flashed chest radiographs. Later on, experiments with eye-trackers showed that abnormalities attracted the attention quite fast but often without further reader actions. Prolonging one’s search seldom leads to newly found abnormalities and may even increase the chance of errors. The problem of reading chest radiographs is therefore not dominated by finding the abnormalities, but by interpreting them. Hypothesis: This suggests that readers could benefit from computer-aided detection (CAD) systems not so much by their ability to prompt potential abnormalities, but more from their ability to `interpret’ the potential abnormalities. In this paper, this hypothesis was investigated by an observer experiment. Experiment: In one condition, the traditional CAD condition, the most suspicious CAD locations were shown to the subjects, without telling them the levels of suspiciousness according to CAD. In the other condition, interactive CAD condition, levels of suspiciousness were given, but only when readers requested them at specified locations. These two conditions focus on decreasing search errors and decision errors, respectively. Results of reading without CAD were also recorded. Six subjects, all non-radiologists, read 223 chest radiographs in both conditions. CAD results were obtained from the OnGuard 5.0 system developed by Riverain Medical (Miamisburg, Ohio). Results: The observer data were analyzed by Location Response Operating Characteristic analysis (LROC). It was found that: 1) With the aid of CAD, the performance is significantly better than without CAD; 2) The performance with interactive CAD is significantly better than with traditional CAD at low false positive rates.
BibTeX:

  @inproceedings{Samu11,
    author = {M.R.M. Samulski and P.R. Snoeren and B. Platel and B. van Ginneken and L. Hogeweg and C. Schaefer-Prokop and N. Karssemeijer},
    title = {Computer-Aided Detection as a Decision Assistant in Chest Radiography},
    booktitle = {Medical Imaging, Proceedings of the SPIE},
    year = {2011},
    volume = {7966},
    pages = {796614}
  }
  
Lesniak, J., Hupse, R., Kallenberg, M., Samulski, M., Blanc, R., Karssemeijer, N. & Székely, G. (2011), “Computer Aided Detection of Breast Masses in Mammography using Support Vector Machine Classification”, In Medical Imaging, Proceedings of the SPIE. Volume 7963, pp. 79631K.

Abstract: The reduction of false positive marks in breast mass CAD is an active area of research. Typically, the problem can be approached either by developing more discriminative features or by employing difierent classifier designs. Usually one intends to find an optimal combination of classifier configuration and small number of features to ensure high classification performance and a robust model with good generalization capabilities. In this paper, we investigate the potential benefit of relying on a support vector machine (SVM) classifier for the detection of masses. The evaluation is based on a 10-fold cross validation over a large database of screenfilm mammograms (10397 images). The purpose of this study is twofold: first, we assess the SVM performance compared to neural networks (NNet), k-nearest neighbor classification (k-NN) and linear discriminant analysis (LDA). Second, we study the classifiers’ performances when using a set of 30 and a set of 73 region-based features. The CAD performance is quantified by the mean sensitivity in 0.05 to 1 false positives per exam on the free-response receiver operating characteristic curve. The best mean exam sensitivities found were 0.545, 0.636, 0.648, 0.675 for LDA, k-NN, NNet and SVM. K-NN and NNet proved to be stable against variation of the featuresets. Conversely, LDA and SVM exhibited an increase in performance when adding more features. It is concluded that with an SVM a more pronounced reduction of false positives is possible, given that a large number of cases and features are available.
BibTeX:

  @inproceedings{Lesn11,
    author = {Jan Lesniak and Rianne Hupse and Michiel Kallenberg and Maurice Samulski and Rémi Blanc and Nico Karssemeijer and Gábor Székely},
    title = {Computer Aided Detection of Breast Masses in Mammography using Support Vector Machine Classification},
    booktitle = {Medical Imaging, Proceedings of the SPIE},
    year = {2011},
    volume = {7963},
    pages = {79631K},
    doi = {http://dx.doi.org/10.1117/12.878140}
  }
  
Radstake, M., Velikova, M., Lucas, P. & Samulski, M. (2011), “Critiquing Knowledge Representation in Medical Image Interpretation using Structure Learning”, In Knowledge Representation for Health-Care (KR4HC). Volume 6512, pp. 56-70. Springer Verlag.

BibTeX:

  @inproceedings{Rads11,
    author = {M. Radstake and M. Velikova and P. Lucas and M. Samulski},
    title = {Critiquing Knowledge Representation in Medical Image Interpretation using Structure Learning},
    booktitle = {Knowledge Representation for Health-Care (KR4HC)},
    publisher = {Springer Verlag},
    year = {2011},
    volume = {6512},
    pages = {56--70}
  }
  
Robben, S., Velikova, M., Lucas, P.J. & Samulski, M. (2011), “Discretisation Does Affect the Performance of Bayesian Networks”, In Research and Development in Intelligent Systems XXVII. , pp. 237-250.

Abstract: In this paper, we study the use of Bayesian networks to interpret breast X-ray images in the context of breast-cancer screening. In particular, we investigate the performance of a manually developed Bayesian network under various discretisation schemes to check whether the probabilistic parameters in the initial manual network with continuous features are optimal and correctly reflect the reality. The classification performance was determined using ROC analysis. A few algorithms perform better than the continuous baseline: best was the entropy-based method of Fayyad and Irani, but also simpler algorithms did outperform the continuous baseline. Two simpler methods with only 3 bins per variable gave results similar to the continuous baseline. These results indicate that it is worthwhile to consider discretising continuous data when developing Bayesian networks and support the practical importance of probabilitistic parameters in determining the network’s performance.
BibTeX:

  @inproceedings{Robb11,
    author = {Robben, Saskia and Velikova, Marina and Lucas, Peter J.F. and Samulski, Maurice},
    title = {Discretisation Does Affect the Performance of Bayesian Networks},
    booktitle = {Research and Development in Intelligent Systems XXVII},
    year = {2011},
    pages = {237--250},
    doi = {http://dx.doi.org/10.1007/978-0-85729-130-1_17}
  }
  
Samulski, M., Hupse, R., Boetes, C., Mus, R.D.M., den Heeten, G.J. & Karssemeijer, N. (2010), “Using computer-aided detection in mammography as a decision support.”, Eur Radiol., Jun, 2010.

Abstract: OBJECTIVE: To evaluate an interactive computer-aided detection (CAD) system for reading mammograms to improve decision making. METHODS: A dedicated mammographic workstation has been developed in which readers can probe image locations for the presence of CAD information. If present, CAD findings are displayed with the computed malignancy rating. A reader study was conducted in which four screening radiologists and five non-radiologists participated to study the effect of this system on detection performance. The participants read 120 cases of which 40 cases had a malignant mass that was missed at the original screening. The readers read each mammogram both with and without CAD in separate sessions. Each reader reported localized findings and assigned a malignancy score per finding. Mean sensitivity was computed in an interval of false-positive fractions less than 10 RESULTS: Mean sensitivity was 25.1% in the sessions without CAD and 34.8% in the CAD-assisted sessions. The increase in detection performance was significant (p = 0.012). Average reading time was 84.7 +/- 61.5 s/case in the unaided sessions and was not significantly higher when interactive CAD was used (85.9 +/- 57.8 s/case). CONCLUSION: Interactive use of CAD in mammography may be more effective than traditional CAD for improving mass detection without affecting reading time.
BibTeX:

  @article{Samu10,
    author = {Maurice Samulski and Rianne Hupse and Carla Boetes and Roel D M Mus and Gerard J den Heeten and Nico Karssemeijer},
    title = {Using computer-aided detection in mammography as a decision support.},
    journal = {Eur Radiol},
    year = {2010},
    doi = {http://dx.doi.org/10.1007/s00330-010-1821-8}
  }
  
Samulski, M., Hupse, A., Boetes, C., den Heeten, G. & Karssemeijer, N. (2009), “Analysis of probed regions in an interactive CAD system for the detection of masses in mammograms”, In Proceedings of SPIE — Volume 7263, Medical Imaging 2009: Image Perception, Observer Performance, and Technology Assessment. Volume 7263(1), pp. 726314. SPIE.

Abstract: Most computer aided detection (CAD) systems for mammographic mass detection display all suspicious regions identified by computer algorithms and are mainly intended to avoid missing cancers due to perceptual oversights. Considering that interpretation failure is recognized to be a more common cause of missing cancers in screening than perceptual oversights, a dedicated mammographic CAD system has been developed that can be queried interactively for the presence of CAD prompts using a mouse click. To assess the potential benefit of using CAD in an interactive way, an observer study was conducted in which 4 radiologists and 6 non-radiologists evaluated 60 cases with and without CAD, to compare the detection performance of the unaided reader with that of the reader with CAD assistance. 20 cases had a malignant mass, and 40 were cancer-free. During the reading sessions we recorded time and probed locations which reveal information about the search strategy and detection process. The purpose of this study is to determine a relation between detection performance and time to first probe of the lesion and to investigate if longer reading times lead to more reports of malignant lesions in lesion-free areas. On average, 65.0% of the malignant lesions were found within 60 seconds and this percentage stabilizes after this period. Results suggest that longer reading time did not lead to more false positives. 74.6% of the reported true positive findings were hit by the first probe, and 93.2% were hit within 5 probes, which may suggest that many of the correctly reported malignant masses were perceived immediately after image onset.
BibTeX:

  @inproceedings{Samu09,
    author = {Samulski, M. and Hupse, A. and Boetes, C. and den Heeten, G. and Karssemeijer, N.},
    title = {Analysis of probed regions in an interactive CAD system for the detection of masses in mammograms},
    booktitle = {Proceedings of SPIE -- Volume 7263, Medical Imaging 2009: Image Perception, Observer Performance, and Technology Assessment},
    journal = {Medical Imaging 2009: Image Perception, Observer Performance, and Technology Assessment},
    publisher = {SPIE},
    year = {2009},
    volume = {7263},
    number = {1},
    pages = {726314},
    url = {http://link.aip.org/link/?PSI/7263/726314/1},
    doi = {http://dx.doi.org/10.1117/12.813391}
  }
  
Velikova, M., Samulski, M., Lucas, P.J.F. & Karssemeijer, N. (2009), “Improved mammographic CAD performance using multi-view information: a Bayesian network framework.”, Phys Med Biol., Mar, 2009. Vol. 54(5), pp. 1131-1147.

Abstract: Mammographic reading by radiologists requires the comparison of at least two breast projections (views) for the detection and the diagnosis of breast abnormalities. Despite their reported potential to support radiologists, most mammographic computer-aided detection (CAD) systems have a major limitation: as opposed to the radiologist’s practice, computerized systems analyze each view independently. To tackle this problem, in this paper, we propose a Bayesian network framework for multi-view mammographic analysis, with main focus on breast cancer detection at a patient level. We use causal-independence models and context modeling over the whole breast represented as links between the regions detected by a single-view CAD system in the two breast projections. The proposed approach is implemented and tested with screening mammograms for 1063 cases of whom 385 had breast cancer. The single-view CAD system is used as a benchmark method for comparison. The results show that our multi-view modeling leads to significantly better performance in discriminating between normal and cancerous patients. We also demonstrate the potential of our multi-view system for selecting the most suspicious cases.
BibTeX:

  @article{Veli09,
    author = {Velikova, Marina and Samulski, Maurice and Lucas, Peter J F and Karssemeijer, Nico},
    title = {Improved mammographic CAD performance using multi-view information: a Bayesian network framework.},
    journal = {Phys Med Biol},
    year = {2009},
    volume = {54},
    number = {5},
    pages = {1131--1147},
    url = {http://dx.doi.org/10.1088/0031-9155/54/5/003},
    doi = {http://dx.doi.org/10.1088/0031-9155/54/5/003}
  }
  
Velikova, M., Samulski, M., Lucas, P.J. & Karssemeijer, N. (2009), “Causal Probabilistic Modelling for Two-View Mammographic Analysis”, In AIME ’09: Proceedings of the 12th Conference on Artificial Intelligence in Medicine. Berlin, Heidelberg., pp. 395-404. Springer-Verlag.

Abstract: Mammographic analysis is a difficult task due to the complexity of image interpretation. This results in diagnostic uncertainty, thus provoking the need for assistance by computer decision-making tools. Probabilistic modelling based on Bayesian networks is among the suitable tools, as it allows for the formalization of the uncertainty about parameters, models, and predictions in a statistical manner, yet such that available background knowledge about characteristics of the domain can be taken into account. In this paper, we investigate a specific class of Bayesian networks–causal independence models–for exploring the dependencies between two breast image views. The proposed method is based on a multi-stage scheme incorporating domain knowledge and information obtained from two computer-aided detection systems. The experiments with actual mammographic data demonstrate the potential of the proposed two-view probabilistic system for supporting radiologists in detecting breast cancer, both at a location and a patient level.
BibTeX:

  @inproceedings{Veli09a,
    author = {Velikova, Marina and Samulski, Maurice and Lucas, Peter J. and Karssemeijer, Nico},
    title = {Causal Probabilistic Modelling for Two-View Mammographic Analysis},
    booktitle = {AIME '09: Proceedings of the 12th Conference on Artificial Intelligence in Medicine},
    publisher = {Springer-Verlag},
    year = {2009},
    pages = {395--404},
    doi = {http://dx.doi.org/10.1007/978-3-642-02976-9_56}
  }
  
Karssemeijer, N., Hupse, A., Samulski, M., Kallenberg, M., Boetes, C. & Heeten, G. (2008), “An Interactive Computer Aided Decision Support System for Detection of Masses in Mammograms”, In IWDM ’08: Proceedings of the 9th international workshop on Digital Mammography. Berlin, Heidelberg., pp. 273-278. Springer-Verlag.

Abstract: A mammographic screening workstation has been developed in which CAD results for mass detection are presented fundamentally different than in current practice. Instead of displaying all CAD findings as prompts the reader can probe image regions for the presence of CAD information. The aim of the system is to help radiologists with decision making rather than avoiding oversight errors. In a preliminary observer study we studied the effect of using the interactive CAD system. Four non-radiologists and two radiologists participated. Each observer read 60 cases two times, once with and once without CAD. The set included 20 cases with subtle cancers that were missed at screening. It was found that performance of the readers increased significantly with interactive use of CAD.
BibTeX:

  @inproceedings{Kars08,
    author = {Karssemeijer, Nico and Hupse, Andrea and Samulski, Maurice and Kallenberg, Michiel and Boetes, Carla and Heeten, Gerard},
    title = {An Interactive Computer Aided Decision Support System for Detection of Masses in Mammograms},
    booktitle = {IWDM '08: Proceedings of the 9th international workshop on Digital Mammography},
    publisher = {Springer-Verlag},
    year = {2008},
    pages = {273--278},
    doi = {http://dx.doi.org/10.1007/978-3-540-70538-3_38}
  }
  
Karssemeijer, N., Samulski, M., den Heeten, G. & Boetes, C. (2008), “Analysis of Observer Performance Based on Probing Patterns in an Interactive CAD System for Mammographic Mass Detection”, In 94th Radiological Society of North America Scientific Assembly and Annual Meeting. Chicago, Illinois, United States

Abstract: PURPOSE
In an interactive CAD system for reading mammograms users may probe regions for CAD information to improve decision making. Probed locations reveal information about the detection process. The purpose of this study is to assess to what extent missed lesions in an observer study were due to perception rather than interpretation failure.
METHOD AND MATERIALS
A mammographic workstation was developed in which readers can use CAD interactively. Instead of displaying all CAD findings as prompts readers can probe image regions for the presence of CAD information using a computer mouse. In an observer study we investigated the effect of this system on detection performance. Seven readers participated in the study, of which two were certified mammographers and five were non-radiologists with mammogram reading skills. The radiologists each read 60 cases including 20 with cancer, the non-radiologists read 120 cases including 40 cancer cases. All cancers selected were missed at the original screening and were retrospectively identified as visible. Cases with only microcalcifications were excluded. Mammograms were digitized from film and cases had up to 8 views when priors were available. Mammograms were read with and without CAD in two different sessions using a counter-balanced study design. The CAD system used was the ImageChecker V8.0 (R2/Hologic) and CAD was available in current and prior mammograms. From logfiles of the reading sessions we could determine locations that were probed during the CAD assisted sessions. Probed locations were correlated with true positive and false negative decisions.
RESULTS
Readers probed locations for CAD information 13.6 times per case on average, (9.9 for the radiologists). On average readers reported 68.8 % of the cancers while 80.8% of them were probed. Performance of the non-radiologists was similar to that of the radiologists. At a false positive recall rate of 10% the mean correct localisation fraction in the CAD assisted sessions was 46%.
CONCLUSION
Results confirm that most missed cancers are due to interpretation failure. On average more than 80% of the cancers were probed for CAD results while only 46% could be detected at a false positive recall rate of 10%.
CLINICAL RELEVANCE/APPLICATION
For further development of CAD it is needed to gain more understanding of perceptual problems in radiology. The proposed approach may serve as an alternative for eye-tracking.
BibTeX:

  @conference{Kars08a,
    author = {Karssemeijer, N. and Samulski, M. and den Heeten, G. and Boetes, C.},
    title = {Analysis of Observer Performance Based on Probing Patterns in an Interactive CAD System for Mammographic Mass Detection},
    booktitle = {94th Radiological Society of North America Scientific Assembly and Annual Meeting},
    year = {2008},
    url = {http://rsna2008.rsna.org/event_display.cfm?em_id=6016239}
  }
  
Karssemeijer, N., Samulski, M., Kallenberg, M., Hupse, A., Boetes, C. & den Heeten, G. (2008), “Effectiveness of an Interactive CAD System for Mammographic Mass Detection”, In 94th Radiological Society of North America Scientific Assembly and Annual Meeting. Chicago, Illinois, United States

Abstract: PURPOSE

To study effectiveness of an interactive CAD system for reading mammograms, in which readers may probe regions for CAD information to improve decision making.

METHOD AND MATERIALS

Interactive use of CAD was studied using a dedicated mammographic workstation. Readers could probe image locations for presence of CAD information using a mouse. If a computer detection was available the CAD region was displayed with a color-coded boundary, ranging from red for regions with a high malignancy rating to green for regions with low ratings. The number of CAD regions that could be activated was set to two false positives per image on average. CAD results were obtained from the ImageChecker V8.0 (Hologic). In an observer study we investigated the effect of this system on detection performance. Seven readers participated in the study, of which two were certified mammographers and five were non-radiologists with mammogram reading experience. The radiologists each read 60 cases including 20 with cancer, the non-radiologists read 120 cases of which 40 had cancer. All cancers selected were missed at screening and were retrospectively identified as visible. Cases with only microcalcifications were excluded. Mammograms were read with and without CAD in two different sessions. Readers reported localized findings with malignancy ratings per finding scored on a continuous scale. Sensitivity, defined as correct cancer localization fraction, was computed per reader as a function of the fraction of normal cases that would be recalled, based on the malignancy ratings. Reading time per case was recorded.

RESULTS

Mean sensitivity in an interval of negative recalls less than 10% was 27.1% in the sessions without CAD and 36.9% in the CAD assisted sessions. Performance increase was significant (p<0.01). Mean sensitivity of the two radiologists was 27.4% and 32.7% resp. without and with CAD, which was comparable to that of the non-radiologists on the same subset of cases. On average, readers probed 13.6 locations per case for CAD information. Average reading time was 105 s/case in the sessions without CAD and was not higher when CAD was used (104 s/case). CONCLUSION Results suggest that detection performance increases with interactive use of CAD without affecting reading time.

CLINICAL RELEVANCE/APPLICATION

Interactive use of CAD in screening mammography is a fundamental change from current practice and may increase effectiveness of the technology for masses.

BibTeX:

  @conference{Kars08b,
    author = {Karssemeijer, N. and Samulski, M. and Kallenberg, M. and Hupse, A. and Boetes, C. and den Heeten, G.},
    title = {Effectiveness of an Interactive CAD System for Mammographic Mass Detection},
    booktitle = {94th Radiological Society of North America Scientific Assembly and Annual Meeting},
    year = {2008},
    url = {http://rsna2008.rsna.org/event_display.cfm?em_id=6020746}
  }
  
Samulski, M. & Karssemeijer, N. (2008), “Matching mammographic regions in mediolateral oblique and cranio caudal views: a probabilistic approach”, In Proceedings of SPIE — Volume 6915, Medical Imaging 2008: Computer-Aided Diagnosis. Volume 6915(1), pp. 69151M. SPIE.

Abstract: Most of the current CAD systems detect suspicious mass regions independently in single views. In this paper we present a method to match corresponding regions in mediolateral oblique (MLO) and craniocaudal (CC) mammographic views of the breast. For every possible combination of mass regions in the MLO view and CC view, a number of features are computed, such as the difference in distance of a region to the nipple, a texture similarity measure, the gray scale correlation and the likelihood of malignancy of both regions computed by single-view analysis. In previous research, Linear Discriminant Analysis was used to discriminate between correct and incorrect links. In this paper we investigate if the performance can be improved by employing a statistical method in which four classes are distinguished. These four classes are defined by the combinations of view (MLO/CC) and pathology (TP/FP) labels. We use distance-weighted k-Nearest Neighbor density estimation to estimate the likelihood of a region combination. Next, a correspondence score is calculated as the likelihood that the region combination is a TP-TP link. The method was tested on 412 cases with a malignant lesion visible in at least one of the views. In 82.4% of the cases a correct link could be established between the TP detections in both views. In future work, we will use the framework presented here to develop a context dependent region matching scheme, which takes the number and likelihood of possible alternatives into account. It is expected that more accurate determination of matching probabilities will lead to improved CAD performance.
BibTeX:

  @inproceedings{Samu08,
    author = {Samulski, Maurice and Karssemeijer, Nico},
    title = {Matching mammographic regions in mediolateral oblique and cranio caudal views: a probabilistic approach},
    booktitle = {Proceedings of SPIE -- Volume 6915, Medical Imaging 2008: Computer-Aided Diagnosis},
    journal = {Medical Imaging 2008: Computer-Aided Diagnosis},
    publisher = {SPIE},
    year = {2008},
    volume = {6915},
    number = {1},
    pages = {69151M},
    url = {http://link.aip.org/link/?PSI/6915/69151M/1},
    doi = {http://dx.doi.org/10.1117/12.769792}
  }
  
Samulski, M. & Karssemeijer, N. (2008), “Linking mass regions in mediolateral oblique and cranio caudal views”, In Proceedings of the 14th ASCI conference. , pp. 214-221.

BibTeX:

  @inproceedings{Samu08a,
    author = {Samulski, Maurice and Karssemeijer, Nico},
    title = {Linking mass regions in mediolateral oblique and cranio caudal views},
    booktitle = {Proceedings of the 14th ASCI conference},
    year = {2008},
    pages = {214--221}
  }
  
Samulski, M., Karssemeijer, N., Boetes, C. & den Heeten, G. (2008), “An Interactive Computer-aided Detection Workstation for Reading Mammograms”, In 94th Radiological Society of North America Scientific Assembly and Annual Meeting. Chicago, Illinois, United States

Abstract: PURPOSE/AIM

To experience the use of an interactive computer-aided decision support system for the detection of mammographic masses. To demonstrate the real-time classification of breast lesions.

CONTENT ORGANIZATION

The idea of using CAD in an interactive way will be explained. Then a case review will be offered, in which participants evaluate a small set of abnormal and normal screening mammograms in two sequential sessions: one without and the other with interactive CAD. Participants are asked to find and rate abnormal masses. At the end of the session, participants can judge their performance and are given the opportunity to review their scores for each case.

SUMMARY

Current computer-aided detection workstations display suspicious mammographic regions identified by computer algorithms as prompts to avoid perceptual oversights. In the presented system the presence of CAD regions can be probed interactively using a mouse click and aid the radiologist with the interpretation of masses. Initial studies suggest that readers may improve their detection performance using CAD in an interactive way.

BibTeX:

  @conference{Samu08b,
    author = {Samulski, M. and Karssemeijer, N. and Boetes, C. and den Heeten, G.},
    title = {An Interactive Computer-aided Detection Workstation for Reading Mammograms},
    booktitle = {94th Radiological Society of North America Scientific Assembly and Annual Meeting},
    year = {2008},
    url = {http://rsna2008.rsna.org/event_display.cfm?em_id=6012408}
  }
  
Velikova, M., Lucas, P.J.F., Ferreira, N., Samulski, M. & Karssemeijer, N. (2008), “A decision support system for breast cancer detection in screening programs”, In Proceeding of the 2008 conference on ECAI 2008. Amsterdam, The Netherlands, The Netherlands., pp. 658-662. IOS Press.

Abstract: The goal of breast cancer screening programs is to detect cancers at an early (preclinical) stage, by using periodic mammographic examinations in asymptomatic women. In evaluating cases, mammographers insist on reading multiple images (at least two) of each breast as a cancerous lesion tends to be observed in different breast projections (views). Most computer-aided detection (CAD) systems, on the other hand, only analyze single views independently, and thus fail to account for the interaction between the views. In this paper, we propose a Bayesian framework for exploiting multi-view dependencies between the suspected regions detected by a single-view CAD system. The results from experiments with real-life data show that our approach outperforms the singleview CAD system in distinguishing between normal and abnormal cases. Such a system can support screening radiologists to improve the evaluation of breast cancer cases.
BibTeX:

  @inproceedings{Veli08,
    author = {Velikova, Marina and Lucas, Peter J. F. and Ferreira, Nivea and Samulski, Maurice and Karssemeijer, Nico},
    title = {A decision support system for breast cancer detection in screening programs},
    booktitle = {Proceeding of the 2008 conference on ECAI 2008},
    publisher = {IOS Press},
    year = {2008},
    pages = {658--662}
  }
  
Velikova, M., Daniels, H. & Samulski, M. (2008), “Partially Monotone Networks Applied to Breast Cancer Detection on Mammograms”, In ICANN ’08: Proceedings of the 18th international conference on Artificial Neural Networks, Part I. Berlin, Heidelberg., pp. 917-926. Springer-Verlag.

Abstract: In many prediction problems it is known that the response variable depends monotonically on most of the explanatory variables but not on all. Often such partially monotone problems cannot be accurately solved by unconstrained methods such as standard neural networks. In this paper we propose so-called MIN-MAX networks that are partially monotone by construction. We prove that this type of networks have the uniform approximation property, which is a generalization of the result by Sill on totally monotone networks. In a case study on breast cancer detection on mammograms we show that enforcing partial monotonicity constraints in MIN-MAX networks leads to models that not only comply with the domain knowledge but also outperform in terms of accuracy standard neural networks especially if the data set is relative small.
BibTeX:

  @inproceedings{Veli08a,
    author = {Velikova, Marina and Daniels, Hennie and Samulski, Maurice},
    title = {Partially Monotone Networks Applied to Breast Cancer Detection on Mammograms},
    booktitle = {ICANN '08: Proceedings of the 18th international conference on Artificial Neural Networks, Part I},
    publisher = {Springer-Verlag},
    year = {2008},
    pages = {917--926},
    doi = {http://dx.doi.org/10.1007/978-3-540-87536-9_94}
  }
  
Velikova, M., Samulski, M., Karssemeijer, N. & Lucas, P. (2008), “Toward Expert Knowledge Representation for Automatic Breast Cancer Detection”, In AIMSA ’08: Proceedings of the 13th international conference on Artificial Intelligence. Berlin, Heidelberg., pp. 333-344. Springer-Verlag.

Abstract: In reading mammograms, radiologists judge for the presence of a lesion by comparing at least two breast projections (views) as a lesion is to be observed in both of them. Most computer-aided detection (CAD) systems, on the other hand, treat single views independently and thus they fail to account for the interaction between the breast views. Following the radiologist’s practice, in this paper, we develop a Bayesian network framework for automatic multi-view mammographic analysis based on causal independence models and the regions detected as suspicious by a single-view CAD system. We have implemented two versions of the framework based on different definitions of multi-view correspondences. The proposed approach is evaluated and compared against the single-view CAD system in an experimental study with real-life data. The results show that using expert knowledge helps to increase the cancer detection rate at a patient level.
BibTeX:

  @inproceedings{Veli08d,
    author = {Velikova, Marina and Samulski, Maurice and Karssemeijer, Nico and Lucas, Peter},
    title = {Toward Expert Knowledge Representation for Automatic Breast Cancer Detection},
    booktitle = {AIMSA '08: Proceedings of the 13th international conference on Artificial Intelligence},
    publisher = {Springer-Verlag},
    year = {2008},
    pages = {333--344},
    doi = {http://dx.doi.org/10.1007/978-3-540-85776-1_28}
  }
  
Samulski, M., Karssemeijer, N., Lucas, P. & Groot, P. (2007), “Classification of mammographic masses using support vector machines and Bayesian networks”, In Proceedings of SPIE — Volume 6514, Medical Imaging 2007: Computer-Aided Diagnosis. Volume 6514(1), pp. 65141J. SPIE.

Abstract: In this paper, we compare two state-of-the-art classification techniques characterizing masses as either benign or malignant, using a dataset consisting of 271 cases (131 benign and 140 malignant), containing both a MLO and CC view. For suspect regions in a digitized mammogram, 12 out of 81 calculated image features have been selected for investigating the classification accuracy of support vector machines (SVMs) and Bayesian networks (BNs). Additional techniques for improving their performance were included in their comparison: the Manly transformation for achieving a normal distribution of image features and principal component analysis (PCA) for reducing our high-dimensional data. The performance of the classifiers were evaluated with Receiver Operating Characteristics (ROC) analysis. The classifiers were trained and tested using a k-fold cross-validation test method (k=10). It was found that the area under the ROC curve (Az) of the BN increased significantly (p=0.0002) using the Manly transformation, from Az = 0.767 to Az = 0.795. The Manly transformation did not result in a significant change for SVMs. Also the difference between SVMs and BNs using the transformed dataset was not statistically significant (p=0.78). Applying PCA resulted in an improvement in classification accuracy of the naive Bayesian classifier, from Az = 0.767 to Az = 0.786. The difference in classification performance between BNs and SVMs after applying PCA was small and not statistically significant (p=0.11).
BibTeX:

  @inproceedings{Samu07,
    author = {Samulski, Maurice and Karssemeijer, Nico and Lucas, Peter and Groot, Perry},
    title = {Classification of mammographic masses using support vector machines and Bayesian networks},
    booktitle = {Proceedings of SPIE -- Volume 6514, Medical Imaging 2007: Computer-Aided Diagnosis},
    journal = {Medical Imaging 2007: Computer-Aided Diagnosis},
    publisher = {SPIE},
    year = {2007},
    volume = {6514},
    number = {1},
    pages = {65141J},
    url = {http://link.aip.org/link/?PSI/6514/65141J/1},
    doi = {http://dx.doi.org/10.1117/12.709679}
  }
  
Samulski, M. (2006), “Classification of Breast Lesions in Digital Mammograms”. Radboud University Nijmegen, June, 2006.

Abstract: Breast cancer is the most common life-threatening type of cancer affecting women in The Netherlands. About 10% of the Dutch women have to face breast cancer in their lifetime. The success of the treatment of breast cancer largely depends on the stage of a tumor at the time of detection. If the size of the invasive cancer is smaller than 20 mm and no metastases are found, chances of successful treatment are high. Therefore, early detection of breast cancer is essential. Although mammography screening is currently the most effective tool for early detection of breast cancer, up to one-fifth of women with invasive breast cancer have a mammogram that is interpreted as normal, i.e., a false-negative mammogram result. An important cause are interpretation errors, i.e., when a radiologist sees the cancer, but classify it as benign. In addition, the number of false-positive mammogram results is quite high, more than half of women who undergo a biopsy actually have breast cancer. To overcome such limitations, Computer-Aided Diagnosis (CAD) systems for automatic classification of breast lesions as either benign or malignant are being developed. CAD systems help radiologists with the interpretation of lesions, such that they refer less women for further examination when they actually have benign lesions. The dataset we used consists of mammographic features extracted by automated image processing algorithms from digitized mammograms of the Dutch screening programme. In this thesis we constructed several types of classifiers, i.e., Bayesian networks and support vector machines, for the task of computer-aided diagnosis of breast lesions. We evaluated the results with receiver operating characteristic (ROC) analysis to compare their classification performance. The overall conclusion is that support vector machines are still the method of choice if the aim is to maximize classification performance. Although Bayesian networks are not primarily designed for classification problems, they did not perform drastically lower. If new datasets are being constructed and more background knowledge becomes available, the advantages of Bayesian networks, i.e., incorporating domain knowledge and modeling dependencies, could play an important role in the future.
BibTeX:

  @mastersthesis{Samu06,
    author = {Samulski, Maurice},
    title = {Classification of Breast Lesions in Digital Mammograms},
    school = {Radboud University Nijmegen},
    year = {2006}
  }
  

Reference list generated by JabRef on 14/06/2010 and exported to wordpress with an export filter originally by Mark Schenk.