Chest Email Content Delivery
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Guest Access | Sign In via User Name/Password
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (13)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Peldschus, K.
Right arrow Articles by Schoepf, U. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Peldschus, K.
Right arrow Articles by Schoepf, U. J.
(Chest. 2005;128:1517-1523.)
© 2005 American College of Chest Physicians

Computer-Aided Diagnosis as a Second Reader*

Spectrum of Findings in CT Studies of the Chest Interpreted as Normal

Kersten Peldschus, MD; Peter Herzog, MD; Susan A. Wood, PhD; Jugesh I. Cheema, MD; Philip Costello, MD and U. Joseph Schoepf, MD

* From the Department of Radiology (Dr. Peldschus and Cheema), Brigham and Women’s Hospital, Harvard Medical School, Boston, MA; Institute of Clinical Radiology (Dr. Herzog), Klinikum Grosshadern, University of Munich, Germany; R2 Technology, Inc. (Dr. Wood), Sunnyvale, CA; and Department of Radiology (Drs. Costello and Schoepf), Medical University of South Carolina, Charleston, SC.

Correspondence to: U. Joseph Schoepf, MD, Department of Radiology, Medical University of South Carolina, 169 Ashley Ave, Charleston, SC 29425; e-mail: schoepf{at}musc.edu


   Abstract
TOP
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Study objectives: To assess the performance of an automated computer-aided detection (CAD) system as a second reader on chest CT studies interpreted as normal at routine clinical interpretation.

Design: Chest CT studies were processed using a prototype CAD system for automated detection of lung lesions. Three experienced radiologists analyzed each CAD finding and confirmed or dismissed the marked image features as lung lesions. Noncalcified, focal lung lesions were classified according to size as being of high (≥ 10 mm), intermediate (5 to 9 mm), or low (≤ 4 mm) significance.

Setting: Two subspecialized academic tertiary referral centers in the United States and Germany.

Patients: Chest CT studies were performed in 100 patients, with results initially reported as normal at clinical double reading. Indications for chest CT were suspected pulmonary embolism (PE) [n = 33], lung cancer screening in a high-risk population (n = 28), or follow-up for a cancer history (n = 39).

Interventions: Reevaluation of all chest CT studies for focal lung lesions with the CAD system as a second reader.

Measurements: Prevalence and spectrum of lung lesions missed at routine clinical interpretation but found by the CAD system.

Results: In 33% (33 of 100 patients), CAD detected significant lung lesions that were not previously reported. Fifty-three significant lesions were detected (mean, 1.6 lesions per case), of which 5 lesions (9.4%) were of high significance, 21 lesions (39.6%) were of intermediate significance, and 27 lesions (50.9%) were of low significance. In the PE group, the lung cancer screening group, and the group with a cancer history, four patients (12.1%), six patients (21.4%), and nine patients (23.1%), respectively, had focal lung lesions of high and/or intermediate significance. The false-positive rate of the CAD system was an average of 1.25 per case (range, 0 to 11).

Conclusions: Significant lung lesions are frequently missed at routine clinical interpretation of chest CT studies but may be detected if CAD is used as an additional reader.

Key Words: cancer screening • lung, neoplasms • lung, nodule • lung neoplasms, diagnosis


   Introduction
TOP
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Diagnoses missed on imaging are a common problem. Of all medical malpractice lawsuits filed against medical imaging professionals, close to 70% are related to the miss of a diagnosis.1 Perceptual errors, which occur when abnormalities that are readily detectable in retrospect are simply "not seen," constitute by far the most common reason why diagnoses are missed and account for approximately 60% of cases.23 The risk of "not seeing" lesions seems greatest in the diagnosis of cancer, and some studies3 have shown rates for this type of error of up to 75%.

CT of the chest is the reference standard for the detection of focal lung disease.4 However, the analytical review of thoracic anatomy and pathology based on many individual transaxial CT sections is demanding. Missing of lesions at thoracic CT is a well-recognized phenomenon,567891011 although the actual magnitude of the problem is difficult to gauge. The risk of missing lung lesions is increased with the introduction of latest generation multidetector-row CT technology, which routinely results in > 200 individual transaxial sections with a slice thickness of ≤ 1 mm for a single chest CT. The information overload associated with reviewing such large, chest CT studies is a cognitive challenge, is time consuming and cumbersome, and greatly increases the risk of lesions being overlooked. Consensus or double readings may help to reduce error1213 but are labor intensive.12

From humble beginnings,14151617 computer-aided diagnosis (CAD) has gained increasing practical importance in medicine in recent years and is successfully applied to cytologic analyses18 and to the detection of cancers of the breast1920 and lung.212223 We designed this study to evaluate whether diagnostic errors can be decreased when CAD algorithms are integrated into clinical imaging as a second reader. We evaluated the prevalence and spectrum of additional findings yielded by CAD in chest CT studies that were read as normal at routine clinical interpretation.


   Materials and Methods
TOP
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
One hundred patients (68 women and 32 men; mean age, 52.9 years; range, 22 to 82 years) who had undergone a CT study of the chest in one of two large academic centers were included. In all cases, the CT study had been interpreted as normal, without suspicious focal lung lesions (inclusion criterion), at routine clinical reading. Three different patient subgroups were evaluated. The first group consisted of 33 consecutive patients (26 women and 7 men; mean age, 44.4 years; range, 22 to 79 years) with suspected pulmonary embolism (PE) but without reported parenchymal focal lung disease. In the second group, there were 28 consecutive individuals (19 women and 9 men; mean age, 58.6 years; range, 37 to 79 years) from a high-risk population (smoking history ≥ 20 pack years) participating in a lung cancer screening program. The third group included 39 consecutive patients (23 women and 16 men; mean age, 55.7 years; range, 27 to 82 years) with a cancer history in full remission undergoing a routine follow-up scan for the following malignant tumor types: prostate cancer (n = 1); ovarian/endometrial cancer (n = 4); Hodgkin disease (n = 4); cancers of unknown primary (n = 6); GI malignancies (n = 7); non-Hodgkin lymphoma (n = 8); and breast cancer (n = 9). The study was performed with institutional review board approval for analysis of CT scan data and records in patients undergoing spiral CT of the chest; informed patient consent was not required.

All studies had been acquired with a 4- or 16-slice multidetector-row spiral CT scanner (Siemens Medical Solutions; Forchheim, Germany). The transaxial images were reconstructed with a thickness varying from 1.25 to 3 mm, resulting in an average number of 241 images per study. All image data were electronically stored at a picture archive and communication system from which they could be retrieved for digital computer-aided image analysis.

All 100 CT studies were processed using a CAD system (ImageCheckerCT; R2 Technology; Sunnyvale, CA). The CAD algorithm is designed to automatically detect and subsequently analyze focal opacities to aid the diagnostic process of a radiologist. It uses a series of volume-centric segmentation steps that delineate normal from abnormal lung tissue. Multiple geometric parameters are calculated for each suspected lung lesion including shape, elongation, size, spiculation, density, and other features. Based on these parameters and following an analytical decision tree, the candidate lesion is given a likelihood rating of representing a lung lesion. If the likelihood rating exceeds a defined threshold for features indicative of a lung lesion, the CAD system deploys a detection mark on the lesion that alerts the human observer to more closely review the marked lung area for the presence of a focal lung lesion. The observer can choose to accept the CAD marked image feature as a focal lung lesion and act on the suspicious finding or dismiss it altogether.

Two experienced radiologists (with 10 years and 7 years of experience in reading thoracic CT, respectively) jointly evaluated and rated each CAD-marked image feature. A third radiologist with 25 years of experience in reading thoracic CT served as an adjudicator for equivocal cases. Each CAD mark was analyzed and was confirmed or dismissed as a lung lesion. Lesions were classified as significant, if they met the criteria of a lung nodule, ie, a circumscribed parenchymal lung lesion that is solid, part solid, or nonsolid in nature and is not associated with atelectasis or adenopathy.24 Significant lung lesions were further classified by size as lesions of high (diameter ≥ 10 mm), intermediate (5 to 9 mm), or low (≤ 4 mm) significance. Lesion location was classified as central or peripheral. CAD-marked image features with a benign calcified pattern or a nonnodular appearance were considered true-positive findings but were categorized as nonsignificant lung changes. CAD-marked image features that by consensus of the readers represented normal thoracic anatomy were considered false-positives.

In cases where significant lung lesions were detected by CAD, the radiology report was amended by an addendum in accordance with the Standard for Communication of the American College of Radiology.2526


   Results
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Within all patients, 285 image features were marked by CAD and were evaluated by the study reading panel. In 33 of the 100 patients (33%), a total of 53 lesions were found that were deemed significant by reader consensus. None of these lesions had been reported at the routine clinical reading. Five of the CAD detected lesions (9.4%) were classified as high significance, 21 lesions (39.6%) were of intermediate significance, and 27 lesions (50.9%) were of low significance. A total of 19 patients (19%) had at least one lesion of high or intermediate significance (Table 1 ).


View this table:
[in this window]
[in a new window]

  		Table 1.. Prevalence and Spectrum of CAD-Detected Findings*

 
In the 33 patients with suspected PE, 20 significant lesions were detected in 12 patients (36.4%). Four patients (12.1%) had focal lung lesions of high significance (one patient) and/or intermediate significance (three patients) [Fig 1 ]. The average number of significant lesions was 0.6 per case.



View larger version (70K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1.. Contrast-enhanced spiral CT study in a 52-year-old woman with acute PE, showing a transaxial section at the level of the left atrium of the heart. Emboli (arrow) are seen in the right intermediate pulmonary artery. In the same section, CAD identified a 6-mm soft-tissue lung nodule (open arrow) that was not reported during the initial clinical readout but was classified as a lesion of intermediate significance by the study readers.

 
In the lung cancer screening group, nine significant lesions were found in 8 of the 28 patients (28.6%). Six patients (21.4%) had findings of high significance (one patient) and/or intermediate significance (five patients) [Fig 2 ]. The average number of significant findings in this group was 0.3 per case.



View larger version (116K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2.. Low-radiation-dose spiral CT study in a 48-year-old man in a lung cancer screening program, showing a transaxial section at the level of the left ventricle (right panel). A 7-mm, slightly spiculated, soft-tissue lung lesion (open arrow), classified as intermediate significance, was missed at clinical interpretation but detected by CAD. A potential explanation why this lesion was missed is its location in the immediate vicinity of lung vessels of similar size, rendering this lesion rather inconspicuous. The location of detected lesions is indicated in a coronal maximum intensity projection of the pulmonary vascular tree (upper left panel). Three-dimensional volume rendering (lower left panel) of the lesion highlighted in green (arrow) displays the spatial relationship of the lesion relative to surrounding vascular structures and enables better differentiation of focal lung lesions and vessels than individual axial sections.

 
In the 39 patients who underwent routine follow-up chest CT for a cancer history, 24 lesions were found in 13 patients (33.3%). Nine patients (23.1%) had focal lung lesions of high significance (three patients) [Fig 3 ] and/or intermediate significance (six patients) [Fig 4 ]. The average number of significant lesions was 0.6 per case.



View larger version (94K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3.. Contrast-enhanced spiral CT study performed for routine follow-up in a 56-year-old woman with a history of breast cancer in full remission, showing a transaxial section at the level of the right phrenicocostal angle (right panel). A yellow line indicates the anatomic level of the displayed section within the volume data set in a coronal maximum intensity projection of the pulmonary vascular tree (left upper panel). An 11-mm, part-solid, slightly spiculated, soft-tissue lung lesion (open arrow), classified as high significance, was missed at clinical interpretation but detected by CAD. Three-dimensional volume rendering (left lower panel) of the lesion highlighted in green (arrow) displays the anatomic location of the lesion in the dorsal right phrenicocostal recess.

 


View larger version (97K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4.. Contrast-enhanced spiral CT study in a 49-year-old woman with a history of colon cancer, showing a transaxial section at the level of the ventricles of the heart (right panel). A 7-mm soft-tissue lung lesion (open arrow), classified as being of intermediate significance, was missed at clinical interpretation but detected by CAD. Similar to Figure 2, the location of the lesion in the immediate vicinity of lung vessels of similar size may have prevented its detection during the initial readout. A yellow line indicates the anatomic level of the displayed section within the volume data set in a coronal maximum intensity projection of the pulmonary vascular tree (left upper panel). Three-dimensional volume rendering (left lower panel) of the lesion highlighted in green (arrow) displays the spatial relationship of the lesion relative to surrounding vascular structures and enables better differentiation of focal lung lesions and vessels than individual axial sections.

 
Most of the missed lesions of high and intermediate significance were located in the central areas of the lung in close proximity to adjacent vessels (Fig 2, 4). Within all candidates, 107 lesions were classified as nonsignificant. Twelve marked lesions were diffusely calcified nodules. Other nonsignificant lesions were mostly apical or subpleural scarring of the lung.

A total of 125 CAD marks were dismissed as false-positive findings. The false-positive rate of the CAD system thus was an average of 1.25 per case (range, 0 to 11). The vast majority of false-positive findings occurred at vessel crossings or vascular structures distorted by image artifacts due to motion (eg, cardiac pulsation).


   Discussion
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
Two types of error can occur during the interpretation of imaging studies: perception error and classification error.23272829 Due to a perception error, a reader simply "does not see" an abnormality that may often be readily visible in retrospect. This can occur, for example, because of a distraction during readout, or because of a particular location of the lesion (eg, a lung nodule adjacent to a vessel of similar size). In a classification error, the observer fails to report a detected, potentially significant lesion because she or he does not attribute adequate clinical importance to the detected lesion.

Although the true magnitude of the problem is difficult to assess, the medical literature is replete with reports on lung lesions that were missed at radiography or CT imaging.567891011 While the missing of lung lesions at imaging is a well-recognized phenomenon, the number of 53 lesions in 33 patients that were not reported at clinical readout in two highly specialized academic departments is surprisingly high. Most of the candidate lesions were < 4 mm and were deemed to be of low significance; however, in five patients (5%) lesions > 1 cm were overlooked.

Significant differences in the prevalence of findings relative to the indication for which chest CT was performed were not seen. Still, slightly more small nodules with a diameter of ≤ 4 mm were found in the group of patients with suspected PE. This may potentially be explained by differences in observer perception, since readers may focus their attention more on vascular structures during the interpretation of studies performed for suspected PE than on nodule detection. Patients in the lung cancer screening group had the lowest per case rate of CAD-detected lesion, likely because in this group of patients the undivided attention of the radiologist is focused on the detection of focal lung lesions.

In this study, we could demonstrate that significant lung lesions missed at the clinical interpretation of chest CT studies can be detected if CAD is used as an additional reader. Use of a CAD system as a second reader thus appears to be able to offset to some degree the inherent variability of human observation by finding oversights that are missed in the original review. Different from a human observer, the CAD system is not affected by distraction or fatigue and always operates at the same performance level. Therefore, the consistency of a CAD system combined with the knowledge and experience of the human observer can have a positive impact on detection accuracy, as demonstrated in this study.

Comparable to other studies,2130 our results show that use of CAD increases the overall accuracy of lesion detection at imaging of the lung. However, detection of additional lung lesions by CAD in a patient with known focal lung disease usually does not alter patient management. CAD detection of lung lesions in a patient whose chest CT study was interpreted as normal, however, as in our study, could potentially result in a significant change in patient management.

There are several aspects of the use of CAD systems that were not addressed by our study design. Computerized systems and neural networks for objectively assessing the likelihood of a lesion of being benign or malignant are actively being investigated. In our study, however, we focused exclusively on the utility of CAD for reducing perception error and not classification error. This latter error is difficult to control for by our retrospective study design. It is possible that some of the 53 lesions that were not reported but were deemed significant for the purpose of this study had indeed been seen, but not considered clinically significant enough to report by the original clinical readers.

It is also difficult to gauge the ultimate significance for patient outcome of the CAD-detected lesions in this study since complete and systematic long-term follow-up or lesion histology are difficult to obtain. However, if a patient is referred for evaluation of focal lung disease, the detection of a lung nodule always bears significance for the ad hoc classification and management of the patient. Potential consequences range from observation and additional follow-up to surgical removal of the lesion, as dictated by the clinical context and patient history. Correct classification and appropriate patient management, however, are jeopardized, if a lesion is missed at diagnostic imaging. In this study, we were able to show that this risk can be mitigated to some extent if CAD is used as an additional safety net.

One hundred twenty-five CAD marks were dismissed as false-positive findings, with an average of 1.25 per case, which is substantially less than with previous iterations of CAD algorithms. However, deployment of a false-positive detection mark in normal thoracic anatomy without pathologic correlate still requires an active decision from the interpreting physician to discard the mark. Dismissal of these detection marks as false-positive findings is usually not very challenging and is accomplished by a single glance of a somewhat experienced reader.

In conclusion, in this study we were able to demonstrate how CAD can be successfully used for reducing perception error in medical imaging. Beneficial integration of such systems into the clinical environment may therefore better ensure that patients with suspected thoracic disease are correctly classified and receive appropriate diagnostic workup and therapy.


   Footnotes
 
Abbreviations: CAD = computer-aided detection; PE = pulmonary embolism

Drs. Peldschus, Costello, and Schoepf had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Dr. Peldschus is supported by the Biomedical Sciences Exchange Program, Hannover, Germany. Dr. Wood is an employee of R2 Technology Inc., Sunnyvale, CA. Dr. Costello is the recipient of a research grant provided by R2 Technology Inc., Sunnyvale, CA. Dr. Schoepf is the recipient of an unrestricted research grant provided by Siemens Corporate Research, Princeton, NJ. The sponsors had no part in the design and conduct of the study or the collection, management, analysis, and interpretation of the data.

Received for publication November 18, 2004. Accepted for publication January 26, 2005.


   References
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Berlin, L, Berlin, J (1995) Malpractice and radiologists in Cook County, IL: trends in 20 years of litigation. AJR Am J Roentgenol 165,781-788[Abstract/Free Full Text]
  2. Renfrew, DL, Franken, EA, Jr, Berbaum, KS, et al Error in radiology: classification and lessons in 182 cases presented at a problem case conference. Radiology 1992;183,145-150[Abstract/Free Full Text]
  3. Berlin, L Malpractice issues in radiology: perceptual errors. AJR Am J Roentgenol 1996;167,587-590[Free Full Text]
  4. Henschke, CI, Yankelevitz, D, Westcott, J, et al Work-up of the solitary pulmonary nodule. American College of Radiology. ACR Appropriateness Criteria. Radiology 2000;215(suppl),607-609
  5. Greene, RE Missed lung nodules: lost opportunities for cancer cure. Radiology 1992;182,8-9[Free Full Text]
  6. Gurney, JW Missed lung cancer at CT: imaging findings in nine patients. Radiology 1996;199,117-122[Abstract/Free Full Text]
  7. White, CS, Romney, BM, Mason, AC, et al Primary carcinoma of the lung overlooked at CT: analysis of findings in 14 patients. Radiology 1996;199,109-115[Abstract/Free Full Text]
  8. Kakinuma, R, Ohmatsu, H, Kaneko, M, et al Detection failures in spiral CT screening for lung cancer: analysis of CT findings. Radiology 1999;212,61-66[Abstract/Free Full Text]
  9. White, CS, Salis, AI, Meyer, CA Missed lung cancer on chest radiography and computed tomography: imaging and medicolegal issues. J Thorac Imaging 1999;14,63-68[ISI][Medline]
  10. Li, F, Sone, S, Abe, H, et al Lung cancers missed at low-dose helical CT screening in a general population: comparison of clinical, histopathologic, and imaging findings. Radiology 2002;225,673-683[Abstract/Free Full Text]
  11. Shah, PK, Austin, JH, White, CS, et al Missed non-small cell lung cancer: radiographic findings of potentially resectable lesions evident only in retrospect. Radiology 2003;226,235-241[Abstract/Free Full Text]
  12. Brown, J, Bryan, S, Warren, R Mammography screening: an incremental cost effectiveness analysis of double versus single reading of mammograms. BMJ 1996;312,809-812[Abstract/Free Full Text]
  13. Gollub, MJ, Panicek, DM, Bach, AM, et al Clinical importance of reinterpretation of body CT scans obtained elsewhere in patients referred for care at a tertiary cancer center. Radiology 1999;210,109-112[Abstract/Free Full Text]
  14. Gorry, GA Prospects for computer-aided diagnosis. N Engl J Med 1969;281,101-102[ISI][Medline]
  15. Sutton, GC How accurate is computer-aided diagnosis? Lancet 1989;2,905-908[ISI][Medline]
  16. Berner, ES, Webster, GD, Shugerman, AA, et al Performance of four computer-based diagnostic systems. N Engl J Med 1994;330,1792-1796[Abstract/Free Full Text]
  17. Kassirer, JP A report card on computer-assisted diagnosis–the grade: C. N Engl J Med 1994;330,1824-1825[Free Full Text]
  18. PRISMATIC Project Management Team.. Assessment of automated primary screening on PAPNET of cervical smears in the PRISMATIC trial. Lancet 1999;353,1381-1385[CrossRef][ISI][Medline]
  19. 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[Abstract/Free Full Text]
  20. 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[Abstract/Free Full Text]
  21. MacMahon, H, Engelmann, R, Behlen, FM, et al Computer-aided diagnosis of pulmonary nodules: results of a large-scale observer test. Radiology 1999;213,723-726[Abstract/Free Full Text]
  22. Armato, SG, III, Li, F, Giger, ML, et al Lung cancer: performance of automated lung nodule detection applied to cancers missed in a CT screening program. Radiology 2002;225,685-692[Abstract/Free Full Text]
  23. Abe, H, MacMahon, H, Engelmann, R, et al Computer-aided diagnosis in chest radiography: results of large-scale observer tests at the 1996–2001 RSNA Scientific Assemblies. Radiographics 2003;23,255-265[Abstract/Free Full Text]
  24. Tuddenham, W Glossary of terms for thoracic radiology: recommendations of the Nomenclature Committee of the Fleischner Society. AJR Am J Roentgenol 1984;43,509-517
  25. American College of Radiology. ACR standard for communication: diagnostic radiology. Standards 2001–2002 2001,3-5 American College of Radiology. Reston, VA:
  26. Orzel, JA, Berlin, L Correction of interpretive errors. AJR Am J Roentgenol 2003;180,1477[Free Full Text]
  27. Berlin, L, Hendrix, RW Perceptual errors and negligence. AJR Am J Roentgenol 1998;170,863-867[Abstract/Free Full Text]
  28. Fitzgerald, R Error in radiology. Clin Radiol 2001;56,938-946[CrossRef][ISI][Medline]
  29. Goddard, P, Leslie, A, Jones, A, et al Error in radiology. Br J Radiol 2001;74,949-951[Abstract/Free Full Text]
  30. Kobayashi, T, Xu, XW, MacMahon, H, et al Effect of a computer-aided diagnosis scheme on radiologists’ performance in detection of lung nodules on radiographs. Radiology 1996;199,843-848[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 The OncologistHome page
M. B. Lobrano
Partnerships in Oncology and Radiology: The Role of Radiology in the Detection, Staging, and Follow-up of Lung Cancer
Oncologist, July 1, 2006; 11(7): 774 - 779.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (13)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Peldschus, K.
Right arrow Articles by Schoepf, U. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Peldschus, K.
Right arrow Articles by Schoepf, U. J.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS