Experiences with computed radiography: can we afford the cost?

In summary, we believe that the CR system adds significantly to our department. The film quality is superb, and the learning curve for radiologists to adapt to the added noise from anatomic structures on chest radiographs is short. The inherent film latitude and flexibility add a new dimension to interpretation and film management. System costs are significantly greater than for traditional film-screen systems. Recurring operational costs are yet to be determined but certainly will be greater with the shorter expected life span of the imaging plates. Film cost savings cannot be a justifiable reason for using a CR system. Film cost savings must await the ultimate conversion to soft imaging through networks and workstations. Technical issues are significant and require structured and specific courses of study and workshops for users before implementation of the CR system. As CR owners, we were frustrated by the lack of published protocols and guides for the CR user. Most of our experiences were learned in practice, and we believed published guides would be useful for both existing users and future buyers. Careful attention to accuracy, histogram selection, positioning, collimation, and patient data input is critical to the success of CR. The persistent lack of a CR information system interface has led to occasional labeling errors and requires added vigilance on the part of both the technologist and the radiologist. Other important user problems we experienced with CR included the RVS. When images are reprinted from the RVS, they can be printed on 10 x 14 inch (26 x 36 cm) film only. Images that were acquired on the larger cassettes are printed on this smaller film, minified by one third from the original size. Minified images are unacceptable to our orthopedic colleagues. Networking problems have still not been resolved, and this resolution remains an important goal. The manipulation and transfer of images without loss of data were major reasons for our switch to electronic imaging and have still not been achieved. Unique image artifacts were encountered, as were new parameters to judge the quality of an image. The final issue is radiation dose. To assign dose reduction as an attribute of CR would be misleading. We have found that in many instances the reverse is true. Many examinations have required more radiation exposure than traditional film-screen techniques. At the least, to harvest the benefits of CR one must significantly alter technical exposure factors and understand how the system operates. The problems we encountered were not minor. Our technologists and radiologists and the Fuji personnel have put in many hours to implement and optimize our system. The image quality is superb, and we continue to work on the important networking and archival goals.

Selective use of image-guided large-core needle biopsy of the breast: accuracy and cost-effectiveness.

OBJECTIVE: We examined the accuracy and cost-effectiveness of large-core needle breast biopsy in a selected group of patients with mammographically detected lesions. MATERIALS AND METHODS: We selectively used large-core needle biopsy to sample breast lesions that were intermediate (neither clearly benign nor clearly malignant) and wire localization biopsy to sample breast lesions that were strongly suggestive of cancer. We compared 2 years' experience using this protocol with the preceding 2 years at the University of Utah Health Sciences Center during which we did only a few large-core needle biopsies on a nonselective basis. RESULTS: Our biopsy rate increased from one biopsy per 36 mammographic screening examinations to one per 26 (p = .001) when the protocol was used. The cost of biopsy per cancer detected decreased from $11,555 to $8356. The specificity of large-core needle biopsy was 98%; the sensitivity based on limited follow-up was 100%. CONCLUSION: Large-core needle biopsy is an accurate and cost-effective method for sampling breast lesions when used in a selective fashion.

Support for biomedical research and its impact on radiology.

Research in medical imaging has experienced substantial growth during the past decade. Still, research is a small fraction of the budget of the typical academic radiology program. Few radiology faculty participate in hypothesis-driven research projects. Funding of research will be more difficult to secure in the future, since clinical subsidies will diminish or disappear, support from industry is decreasing, and funds from private foundations and philanthropists are not likely to increase. Support from the NIH will probably remain about level in constant dollars. In response to these constraints, radiology will have to be both more creative and more opportunistic to tap the limited remaining resources of research support. An excellent compilation of some major resources was recently published by Williams and Holden (9). Efforts of the Conjoint Committee will continue to be critical for continuing support of the LDRR, encouraging the allocation of intramural and extramural resources of the NCI to medical imaging, guiding the development of the American Academy of Radiologic Research, providing research training opportunities for physicians and scientists in radiology, and leading the research effort in medical imaging in general (10). Within individual institutions and departments, imaging research must continue to be acknowledged as a priority despite increasing pressures to generate clinical revenue. Enhanced efforts are warranted to nurture the research interests of younger faculty and selected residents and fellows, including pairing them with research mentors and providing them with opportunities to develop skills in areas such as research design, statistical analysis, and evaluative techniques. The long-term well-being of radiology and its important contributions to patient care are dependent on its continued investment in research and development.

Lymphoproliferative disorders of the lung: histopathology, clinical manifestations, and imaging features.

The lymphoproliferative disorders represent a spectrum of lymphoid abnormalities that can involve the chest. Plasma cell granuloma, pseudolymphoma, posttransplantation lymphoproliferative disorders, lymphoid interstitial pneumonia, and lymphomatoid granulomatosis involve the pulmonary parenchyma, whereas Castleman's disease, infectious mononucleosis, and angioimmunoblastic lymphadenopathy with dysproteinemia involve intrathoracic lymph nodes. Recent immunohistochemical techniques give us a better understanding of the lymphoproliferative disorders. Clinical and radiologic features often allow differentiation of the lymphoproliferative disorders from the more common aggressive lymphomas.

The diagnosis and staging of primary lung cancer.

The controversy surrounding the diagnosis and staging of the patient with primary lung cancer is reviewed in this article. Three basic algorithmic approaches to the problem are presented, with recommendations for both the theoretic and practical approaches to staging.

Imaging in gynecologic malignancies.

Cross-sectional imaging is often a useful complement to clinical examination in patients with gynecologic malignancies. Patients with ovarian cancer will benefit less from the use of computed tomography and magnetic resonance imaging (MRI); however, extensive intraperitoneal disease may be evaluated and followed with either technology. Endometrial cancer is best approached by staging with contrast-enhanced MRI. The more advanced the disease is, the more information should be available from MRI imaging. By contrast, patients with cervical cancer are assessed best using noncontrast-enhanced MRI as far as imaging of the primary tumor is concerned. Extracervical extension is defined better with contrast enhancement and varied pulse sequences. Radiation changes and tumor response also can be assessed with MRI imaging as a complement to clinical examination.

Current applications of imaging procedures in the patient with lung cancer.

The primary role of imaging procedures in the patient with lung cancer should be focused on staging and follow-up challenges. The role of imaging procedures in the detection of the patient at risk for primary lung cancer remains limited and cannot be recommended at present. There is no significant difference between the yield of CT and MR in this patient group, with the possible exception of a more specific role for MR when questions are raised concerning hilar lymph node involvement and mediastinal compartmental invasion. The main role of cross-sectional imaging techniques should be in the avoidance of unnecessary surgical procedures, identifying the unresectable patient prior to exploratory thoracotomy. It should be emphasized that all radiographic abnormalities are non-specific and must be histologically verified before presuming that an abnormal lymph node or large adrenal gland contains metastatic lung cancer.

The applications of imaging in lung cancer.

The applications of imaging to the lung cancer patient have become more focused recently. Screening of high risk patients is not recommended even though intuition and clinical judgment prevail in practice to justify the use of chest radiographs in this patient category. Cross-sectional imaging procedures should be tailored to the staging process in the individual with a large central primary or to confirm an abnormality noted on the chest radiographs. The patient at high risk for thoracotomy is generally also subjected to radiologic staging. The radiologic staging process is reviewed and critiqued, emphasizing our role in identifying the disease sites that would suggest nonrespectability.

State-of-the-art assessment. Diagnostic oncologic imaging.

This review has summarized the status of organ site cancer imaging as applied to tumor detection, staging, and posttreatment follow-up. More general questions which have not been addressed include those related to the following: (1) the problem of providing more adequate training of radiologists in the specific challenge of cancer imaging; (2) how to increase the awareness of oncologists as to the specific indications and applications of tumor imaging procedures and enhance joint communication between radiologists and clinicians in the planning of the imaging procedures; and (3) how to stimulate the radiology and oncology communities to establish imaging standards and recommended procedures for specific tumor imaging challenges. Hopefully, an appreciation of the complex challenge of cancer imaging will result from these and subsequent discussions.

Effects of improved contrast on lung-nodule detection. A clinical ROC study.

We evaluated the effects of unsharp masking and highly efficient scatter rejection on film-screen chest radiographs of cancer patients. Unsharp masking significantly improved the detectability of lung nodules and visibility of anatomic structures in poorly penetrated areas of the chest. Highly efficient scatter rejection by an improved antiscatter grid provided only modest additional benefits. The study supports the conclusion that nodule detection in poorly penetrated areas on conventional chest radiographs is limited primarily by display contrast, whereas in the well-penetrated lung fields it is limited primarily by confusing background structures, rather than inadequate contrast. A method for analyzing clinical nodule detection data by transforming the FROC data to ROC coordinates also is demonstrated.

Non-Hodgkin's lymphoma of the head and neck: CT evaluation of nodal and extranodal sites.

Forty-five patients with non-Hodgkin's lymphoma (NHL) of the extracranial head and neck who had undergone CT as part of their evaluation were reviewed to assess the impact of CT on clinical management. The sites of tumor deposition were subdivided by location: I, nodal; II, extranodal, lymphatic (Waldeyer's ring); and III, extranodal, extralymphatic (orbit, sinonasal, deep facial spaces, mandible, salivary gland, skin, and larynx). The CT appearance of NHL in each of the three locations was analyzed for characteristic CT signatures. Nodal NHL was suspected when CT showed multiple, large, homogeneous lymph nodes, often in unusual nodal chains of the head and neck. Extranodal, lymphatic NHL of Waldeyer's ring was indistinguishable from squamous cell carcinoma of this area unless synchronous tumor deposit in an extranodal, extralymphatic location was also present. When NHL was in nodes and/or Waldeyer's ring, CT-derived information was of limited clinical value since treatment was unfocused (chemotherapy and/or large-field radiotherapy). The CT appearances of extranodal, extralymphatic NHL was generally not distinguishable from other malignancies of these areas. However, CT-derived information regarding deep-tissue tumor size and extent was critical to planning the radiotherapy ports.