Emeritus Faculty - University Medical Line, Radiology
Assistant Professor of Radiology, University of California, SOM (1979 - 1982)
Associate Professor of Radiology;Chief of Section-CT Body Scanning & GI Radiology, University of California, SOM (1982 - 1984)
Professor of Radiology, Chief, Department of Radiology, San Francisco General Hospital (1984 - 1989)
Vice-Chairman, Department of Radiology, University of California, SOM (1984 - 1989)
Chief of Medical Staff, San Francisco General Hospital (1987 - 1988)
Professor and Chairman, Department of Radiology, University of Pittsburgh Medical Center (1989 - 1992)
Director, Abdominal Imaging, Department of Radiology, University of Pittsburgh Medical Center (1992 - 2008)
Chief, Quality Process and Improvement, Department of Radiology, University of Pittsburgh Medical Center (2008 - 2008)
Professor and Associate Chair for Education, Department of Radiology, Stanford University Medical Center (2008 - Present)
Honors & Awards
Blue Ribbon Award-Best Scientific Lecture/Exhibit, American Urological Association, Boston, Massachusetts (1981)
Cum Laude Award-Radiologic-Pathologic Correlation in Liver Cirrhosis, (Society of Computed Body Tomography) (1993)
Magna cum laude award Spectrum of imaging findings in hepatic cirrhosis, Radiological Society of North America (1993)
Presidential Service Award, Pittsburgh Roentgen Society (1995)
Certificate of Merit Citation, RSNA (1999)
Editor's Recognition Award with Distinction, Radiology (1996 2000)
Editor's Certificate of Recognition, Radiographics (2000)
Michael P. Federle Mentorship Award (First Recipient), University of Pittsburgh School of Medicine (2207)
Teaching Division of the Year, Department of Radiology, University of Pittsburgh School of Medicine (2008)
Walter B. Cannon Medal, The Society of Gastrointestinal Radiologists (2010)
N/A, Harvard University, Program for Chiefs, Clin. Svcs. (1989)
Clinical Instructor, UCSF School of Medicine, CT Body/Scanning/GI (1979)
Resident, U of Cincinnati Hosp., Radiology (1978)
Intern, U of Cincinnati Hosp., Internal Medicine (1975)
MD, Georgetown University, Medicine (1974)
Independent Studies (6)
- Directed Reading in Radiology
RAD 299 (Aut, Win, Spr, Sum)
- Early Clinical Experience in Radiology
RAD 280 (Aut, Win, Spr, Sum)
- Graduate Research
RAD 399 (Aut, Win, Spr, Sum)
- Medical Scholars Research
RAD 370 (Aut, Win, Spr, Sum)
- Readings in Radiology Research
RAD 101 (Aut, Win, Spr, Sum)
- Undergraduate Research
RAD 199 (Aut, Win, Spr, Sum)
- Directed Reading in Radiology
Hepatocellular Carcinoma and Intrahepatic Cholangiocarcinoma: Imaging for Diagnosis, Tumor Response to Treatment and Liver Response to Radiation.
Seminars in radiation oncology
2018; 28 (4): 267–76
Hepatocellular carcinoma (HCC) and intrahepatic cholangiocarcinoma (ICC) comprise the majority of primary liver cancers. Both HCC and ICC have characteristic imaging appearances on multiphase computed tomography (CT) and magnetic resonance imaging (MRI). Several locoregional therapies, including radiation therapy, are used to treat unresectable disease and residual or recurrent tumor. The tumor response following locoregional therapies has variable imaging manifestations. Focal liver reaction, the imaging changes of the liver following radiation treatment, should be recognized and not mistaken for tumor. We review the diagnostic imaging of HCC and ICC, imaging of tumor response to treatment, and imaging of the liver response to radiation.
View details for PubMedID 30309637
Contrast media for fluoroscopic examinations of the GI and GU tracts: current challenges and recommendations
2017; 42 (1): 90-100
One of the significant challenges facing radiologists who perform and interpret studies of the gastrointestinal and genitourinary systems have been periodic interruptions in the availability of barium and iodinated contrast media specially formulated for gastrointestinal (GI) and genitourinary (GU) studies. These interruptions are due to the US Food and Drug Administration's recent requirement for more stringent documentation of the safety and efficacy of contrast media and the consolidation among contrast manufacturers. Therefore, radiologists may be required to recommend an alternative means of evaluation, such as computed tomography, magnetic resonance, or endoscopy, or they may need to substitute a different formulation of a contrast agent not specifically developed for GI or GU use, for example the utilization of an agent designed and marketed for vascular use. This article reviews the current status of fluoroscopic contrast media, and provides suggestions and recommendations for the optimal and alternative use of contrast media formulations.
View details for DOI 10.1007/s00261-016-0861-1
View details for Web of Science ID 000392291700009
View details for PubMedID 27503380
Aortoenteric fistulas: spectrum of CT findings
2013; 38 (2): 367-375
This article reviews the causes of aortoenteric fistulas, diagnostic options, and important CT findings.Aortoenteric fistula, a rare but potentially fatal entity, presents a significant challenge to radiologists in diagnosis, largely because of its subtle and nonspecific imaging findings. These fistulas can be divided into primary and secondary forms, depending on the presence or absence of prior aortic reconstructive surgery, but the secondary form is more common. Typical CT findings, which can overlap with those seen in perigraft infection, aortitis, infected/mycotic aneurysms, perianeurysmal fibrosis, and the immediate post-operative period after placement of a graft, include: Effacement of the fat planes around the aorta, perigraft fluid/soft tissue thickening, ectopic gas, tethering of adjacent thickened bowel loops towards the aortic graft, and in rare cases, extravasation of contrast from the aorta into the involved segment of bowel.
View details for DOI 10.1007/s00261-012-9873-7
View details for Web of Science ID 000316144100019
View details for PubMedID 22366854
Imaging Manifestations of Abdominal Fat Necrosis and Its Mimics
2011; 31 (7): 2021-2034
Intraabdominal fat is a metabolically active tissue that may undergo necrosis through a number of mechanisms. Fat necrosis is a common finding at abdominal cross-sectional imaging, and it may cause abdominal pain, mimic findings of acute abdomen, or be asymptomatic and accompany other pathophysiologic processes. Common processes that are present in fat necrosis include torsion of an epiploic appendage, infarction of the greater omentum, and fat necrosis related to trauma or pancreatitis. In addition, other pathologic processes that involve fat may be visualized at computed tomography, including focal lipohypertrophy, pathologic fat paucity (lipodystrophies), and malignancies such as liposarcoma, which may mimic benign causes of fat stranding. Because fat necrosis and malignant processes such as liposarcoma and peritoneal carcinomatosis may mimic one another, knowledge of a patient's clinical history and prior imaging studies is essential for accurate diagnosis.
View details for DOI 10.1148/rg.317115046
View details for PubMedID 22084185
Radiologists' performance in the diagnosis of acute intestinal ischemia, using MDCT and specific CT findings, using a variety of CT protocols
2011; 18 (5): 385–94
The aim of this study was to evaluate the performance of radiologists in the diagnosis of acute intestinal ischemia using specific multi-detector CT findings. The abdominal CT scans of 90 patients were retrospectively reviewed by three radiologists: an abdominal imaging specialist, an experienced general radiologist, and a senior resident. Forty-seven patients had surgically proven intestinal ischemia and comprised the case group, while 43 patients had no evidence of intestinal ischemia at surgery and comprised the control group. Images were reviewed in a random and blinded fashion. Radiologists' performance in diagnosing bowel ischemia from other bowel pathologies was evaluated. The sensitivity, specificity, and accuracy for diagnosing bowel ischemia were 89%, 67%, and 79% for the abdominal imager; 83%, 67%, and 76% for the general radiologist; and 66%, 83%, and 74% for the senior resident, respectively. The calculated kappa value for inter-observer agreement regarding the presence of bowel ischemia was 0.79. CT findings that significantly distinguished bowel ischemia from other bowel pathologies were decreased or absent bowel wall enhancement, filling defect in the superior mesenteric artery, small bowel pneumatosis, and gas in the portal veins or superior mesenteric vein. For most of these signs, there was good inter-observer agreement. Radiologists' performance in diagnosing bowel ischemia is good, but lower than previously reported since a significant amount of cases are evaluated using a suboptimal CT technique. Radiologists' experience and expertise have an important impact on their performance.
View details for DOI 10.1007/s10140-011-0965-4
View details for Web of Science ID 000217344000004
View details for PubMedID 21655965
Septic Thrombophlebitis of the Portal Venous System: Clinical and Imaging Findings in Thirty-Three Patients
DIGESTIVE DISEASES AND SCIENCES
2011; 56 (7): 2179-2184
Our purpose was to review the clinical and imaging findings in a series of patients with septic thrombophlebitis of the portal venous system in order to define criteria that might allow more confident and timely diagnosis.This is a retrospective case series. The clinical and imaging features were analyzed in 33 subjects with septic thrombophlebitis of the portal venous system.All 33 patients with septic thrombophlebitis of the portal venous system had pre-disposing infectious or inflammatory processes. Contrast-enhanced CT studies of patients with septic thrombophlebitis typically demonstrate an infectious gastrointestinal source (82%), thrombosis (70%), and/or gas (21%) of the portal system or its branches, and intrahepatic abnormalities such as a transient hepatic attenuation difference (THAD) (42%) or abscess (61%).Septic thrombophlebitis of the portal system is often associated with an infectious source in the gastrointestinal tract and sepsis. Contrast-enhanced CT demonstrates an infectious gastrointestinal source, thrombosis or gas within the portal system or its branches, and intrahepatic abnormalities such as abscess in most cases. We report a THAD in several of our patients, an observation that was not made in prior reports of septic thrombophlebitis.
View details for DOI 10.1007/s10620-010-1533-6
View details for Web of Science ID 000291481800038
View details for PubMedID 21221797
Chemoembolization for Unresectable HepatoCellular Carcinoma in Patients with or without Portal Vein Thrombosis
2010; 58 (104): 1375-1381
Chemoembolization in the presence of portal vein thrombosis (PVT) is thought to be hazardous and such patients are regarded as prognostically poor.One hundred sixty two patients with biopsy-proven and unresectable HCC were treated with chemoembolization (TACE). We retrospectively analyzed these patients according to presence or absence of PVT.We found the 2 groups were similar with respect to tumor and liver parameters. Survival was 22.3 months in responders and 6.6 months in non-responders, p < or = 0.0001. Patients with PVT who responded had a 24 mo median survival and without PVT who responded had a 30 month median survival. These 2 groups were thus similar, p = 0.817. Patients who did not respond had a survival of only 5 and 7 months, for presence or absence of PVT, respectively. Responders had significantly lower blood AFP and GGTP levels, lower DCP levels and better liver function.In this series, response was a major determinant of survival, regardless of the presence or absence of PVT. Thus, branch PVT was not a contraindication to chemoembolization.
View details for Web of Science ID 000287685900010
Managing Incidental Findings on Abdominal CT: White Paper of the ACR Incidental Findings Committee
JOURNAL OF THE AMERICAN COLLEGE OF RADIOLOGY
2010; 7 (10): 754-773
As multidetector CT has come to play a more central role in medical care and as CT image quality has improved, there has been an increase in the frequency of detecting "incidental findings," defined as findings that are unrelated to the clinical indication for the imaging examination performed. These "incidentalomas," as they are also called, often confound physicians and patients with how to manage them. Although it is known that most incidental findings are likely benign and often have little or no clinical significance, the inclination to evaluate them is often driven by physician and patient unwillingness to accept uncertainty, even given the rare possibility of an important diagnosis. The evaluation and surveillance of incidental findings have also been cited as among the causes for the increased utilization of cross-sectional imaging. Indeed, incidental findings may be serious, and hence, when and how to evaluate them are unclear. The workup of incidentalomas has varied widely by physician and region, and some standardization is desirable in light of the current need to limit costs and reduce risk to patients. Subjecting a patient with an incidentaloma to unnecessary testing and treatment can result in a potentially injurious and expensive cascade of tests and procedures. With the participation of other radiologic organizations listed herein, the ACR formed the Incidental Findings Committee to derive a practical and medically appropriate approach to managing incidental findings on CT scans of the abdomen and pelvis. The committee has used a consensus method based on repeated reviews and revisions of this document and a collective review and interpretation of relevant literature. This white paper provides guidance developed by this committee for addressing incidental findings in the kidneys, liver, adrenal glands, and pancreas.
View details for DOI 10.1016/j.jacr.2010.06.013
View details for Web of Science ID 000208363200007
Xanthogranulomatous cholecystitis: Diagnostic performance of CT to differentiate from gallbladder cancer
EUROPEAN JOURNAL OF RADIOLOGY
2010; 74 (3): E80-E84
To retrospectively evaluate CT findings of xanthogranulomatous cholecystitis (XGC) and to measure diagnostic performance for distinguishing it from gallbladder (GB) cancer.Our institutional review board approved this retrospective study. Three blinded radiologists, first independently and then in consensus, retrospectively evaluated postcontrast CT images of 35 patients with histopathologically proved XGC and GB cancer, all of whom subsequently had cholecystectomy. These included 18 patients with XGC (13 male, 5 female; age range, 35-84, mean 63 years) and 17 with GB cancer (6 male, 11 female; age range, 45-95, mean 69). Differences in CT findings between XGC and GB cancer and diagnostic performances for each CT finding were calculated. Sensitivity, specificity, and accuracy were calculated for each radiologist and observer performance was also determined by receiver-operating-characteristic curve analysis.Five CT findings showed significant differences between XGC and GB cancer. Sensitivity, specificity, and accuracy of each finding for the differentiation of XGC were 89%, 65%, 77% with diffuse GB wall thickening, 67%, 82%, 74% with a continuous mucosal line, 61%, 71%, 66% with intra-mural hypo-attenuated nodules, 72%, 77%, 74% with absence of macroscopic hepatic invasion, and 67%, 71%, 69% with absence of intra-hepatic bile duct dilatation, respectively. When at least three of these five CT findings were observed in combination, sensitivity, specificity, and accuracy were 83%, 100% and 91%, respectively. Sensitivities, specificities and Az values for the differentiation of XGC from GB cancer were 83%, 88%, 0.94 for reader 1, 78%, 88%, 0.93 for reader 2, and 78%, 82%, 0.84 for reader 3.The combination of three of the five CT findings that are common with XGC can provide excellent accuracy for the differentiation of XGC and GB cancer.
View details for DOI 10.1016/j.ejrad.2009.04.017
View details for Web of Science ID 000279350800035
Distinguishing clinical and imaging features of nodular regenerative hyperplasia and large regenerative nodules of the liver
2009; 64 (12): 1190-1195
Nodular regenerative hyperplasia (NRH) and large regenerative nodules (LRN) are distinct types of hepatocellular nodules that have been confused in the radiology literature. However, distinction is critical because their clinical significance is quite different. Our purpose was to review the clinical and imaging findings in a series of patients with NRH and LRN in order to identify distinguishing clinical and imaging features.This was a retrospective case series. The clinical and imaging features were compared in 36 patients with pathological proof of NRH and 23 patients with pathological evidence of LRN.NRH and LRN have different predisposing factors and imaging findings. NRH is often associated with organ transplantation, myeloproliferative disease, or autoimmune processes. Livers with NRH typically do not have enhancing nodules; none of the present patients with NRH had enhancing liver masses. In contrast, LRN are often associated with Budd-Chiari syndrome. Enhancing liver masses were noted in 19 (83%) of the 23 patients with LRN. The p values for the comparisons were less than 0.001 for both enhancing liver masses and hepatic vein thrombosis.NRH and LRN can have distinct clinical presentations and imaging appearances. LRN often result in enhancing liver nodules, whereas NRH usually does not. Clinical and imaging information enables the distinction of LRN and NRH in many cases.
View details for DOI 10.1016/j.crad.2009.07.015
View details for Web of Science ID 000273305900007
View details for PubMedID 19913129
Spontaneous Abdominal Hemorrhage: Causes, CT Findings, and Clinical Implications
AMERICAN JOURNAL OF ROENTGENOLOGY
2009; 193 (4): 1077-1087
The purpose of this article is to present the most common causes of spontaneous abdominal hemorrhage and to review the CT findings that are important in establishing the correct diagnosis and in guiding appropriate therapy.Knowledge of the common CT manifestations of various causes of spontaneous abdominal hemorrhage allows their accurate diagnosis and has a direct impact on clinical decision making.
View details for DOI 10.2214/AJR.08.2231
View details for Web of Science ID 000270033300023
View details for PubMedID 19770332
Liver Lesions With Hepatic Capsular Retraction
SEMINARS IN ULTRASOUND CT AND MRI
2009; 30 (5): 426-435
Retraction of the liver capsule may be associated with a diverse spectrum of benign and malignant etiologies. The more common causes include focal confluent fibrosis in cirrhotic livers, cholangiocarcinoma, and treated liver tumors, such as hepatocellular carcinoma, metastases, and lymphoma. Less common etiologies include primary sclerosing cholangitis, epithelioid hemangioendothelioma, hepatic hemangioma, solitary fibrous tumor of the liver, and hepatic inflammatory pseudotumor. Hepatic capsular retraction may also result from iatrogenic and noniatrogenic trauma. Due to the diversity and different nature of the various etiologies associated with this sign, it is important that radiologists be familiar with the characteristic features of these abnormalities, to avoid misdiagnosis that may adversely affect the therapeutic approach. It is also important to know that, contrary to some reports, hepatic capsular retraction is not a sign of malignant disease. The purpose of this article is to familiarize readers with the spectrum of benign and malignant etiologies of this sign and to point out additional computed tomographic findings that may allow confident diagnosis of the specific hepatic lesion responsible for the capsular retraction. The hepatic capsular and subcapsular regions may be affected by focal or diffuse pathologies affecting the liver. This hepatic area is more prone to be involved in various malignant and benign diseases due to several factors: the negative subdiaphragmatic pressure that may draw infected material and malignant cells toward the diaphragm, the perihepatic ligaments connecting the liver capsule with adjacent viscera, forming a direct root of dissemination, and the systemic blood inflow that supplies this region in addition to the portal and hepatic arterial blood flow. This is the reason for the multiple pathologic conditions and pseudolesions that occur at the hepatic capsular and subcapsular regions.
View details for DOI 10.1053/j.sult.2009.06.002
View details for Web of Science ID 000270442900006
View details for PubMedID 19842567
Imaging Approach for Evaluation of Focal Liver Lesions
CLINICAL GASTROENTEROLOGY AND HEPATOLOGY
2009; 7 (6): 624-634
Focal liver lesions are common in the general population. Radiology (imaging) plays a pivotal role for the diagnosis, staging, treatment planning, and follow-up of focal liver lesions. To maximize lesion detection and characterization, imaging needs to be performed with appropriate equipment by using protocols carefully designed on the basis of the underlying clinical context. In addition, the decision of an imaging modality cannot be based on the diagnostic accuracy of an imaging test solely but must also consider patient safety and cost-effectiveness.
View details for DOI 10.1016/j.cgh.2009.03.024
View details for Web of Science ID 000267189400006
View details for PubMedID 19348962
CT Hypotension Complex (Shock Bowel) Is Not Always Due to Traumatic Hypovolemic Shock
AMERICAN JOURNAL OF ROENTGENOLOGY
2009; 192 (5): W230-W235
The purpose of our study was to review the clinical and CT findings in a substantial series of 41 patients with the shock bowel sign to determine if there is an association between shock bowel (and other CT signs of hypotension) and conditions other than post-traumatic hypovolemic shock.The shock bowel sign and the CT hypotension complex are frequently associated with hypotension from causes other than trauma-induced hypovolemic shock, such as severe head or spine injury, cardiac arrest, septic shock, bacterial endocarditis, and diabetic ketoacidosis. Other elements of the CT hypotension complex such as flattening of the inferior vena cava and aorta, abnormal pancreatic enhancement and peripancreatic fluid, and hypoperfusion of the spleen and liver are variably associated with shock bowel whether due to posttraumatic hypovolemia or other causes of hypotension. The CT hypotension complex (shock bowel) has important prognostic and therapeutic implications and can probably be distinguished from bowel trauma and other forms of bowel injury in most cases.
View details for DOI 10.2214/AJR.08.1474
View details for Web of Science ID 000265387300049
View details for PubMedID 19380528
Focal Confluent Fibrosis in Cirrhotic Liver: Natural History Studied with Serial CT
AMERICAN JOURNAL OF ROENTGENOLOGY
2009; 192 (5): 1341-1347
The objective of this study was to assess the long-term natural history of focal confluent fibrosis in cirrhotic liver with CT.Two radiologists retrospectively reviewed in consensus 118 liver CT examinations in 26 patients (19 men, seven women; age range, 32-68 years; mean age, 50 years) performed over approximately 6 years. Helical CT scans were obtained before and 30-35 and 65-70 seconds after injection of 125-150 mL of contrast medium at a rate of 4-5 mL/s. Proof of cirrhosis was based on liver transplantation (n = 6), biopsy (n = 9), or imaging findings (n = 11). The number, location, and attenuation of fibrotic lesions and presence of trapped vessels were evaluated. Variation of hepatic retraction associated with the development of focal confluent fibrosis lesions was assessed using the ellipsoid volume formula and an arbitrary retraction index.Each radiologist identified 41 focal confluent fibrosis lesions. All lesions were identified by both radiologists. Twelve patients (46%) had a single lesion, 13 (50%) had two lesions, and one (4%) had three lesions. Thirty-four (83%) of 41 lesions were located in segment IV, VII, or VIII. Thirty-two lesions (78%) were hypoattenuating on unenhanced images, 25 lesions (61%) were hypoattenuating on hepatic arterial phase images, and 20 lesions (49%) were isoattenuating on portal venous phase images. Seven lesions (17%) were or became hyperattenuating at follow-up on portal venous phase images. Trapped vessels were found in six lesions (15%). The retraction index showed a significant increase over time (r = 0.423, p < or = 0.0001).The degree of capsule retraction associated with focal confluent fibrosis evolves with time and relates to the natural evolution of cirrhosis.
View details for DOI 10.2214/AJR.07.2782
View details for Web of Science ID 000265387300028
View details for PubMedID 19380559
Using contrast-enhanced helical CT to visualize arterial extravasation after blunt abdominal trauma: Incidence and organ distribution
AMERICAN JOURNAL OF ROENTGENOLOGY
2002; 178 (1): 17-20
We evaluated the incidence and organ distribution of arterial extravasation identified using contrast-enhanced helical CT in patients who had sustained abdominal visceral injuries and pelvic fractures after blunt trauma.Five hundred sixty-five consecutive patients from four level I trauma centers who had CT scans showing abdominal visceral injuries or pelvic fractures were included in this series. The presence or absence of arterial extravasation, as well as the anatomic sites of arterial extravasation, was noted. We obtained clinical follow-up data, including surgical or angiographic findings.In our series, 104 (18.4%) of 565 patients had arterial extravasation. Of the 104 patients, 81 (77.9%) underwent surgery, embolization, or both. The combined rate of surgery or embolization in patients with arterial extravasation was statistically higher than expected at all four institutions (p <0.001). The spleen was the most common organ injured, occurring in 277 (49.0%) of 565 patients, and arterial extravasation occurred in 49 (17.7%) of 277 patients with splenic injury. Several other visceral injuries were associated with arterial extravasation, including hepatic, renal, adrenal, and mesenteric injuries.Based on the limited reports of arterial extravasation in the nonhelical CT literature, the percentage (18%) of clinically stable patients in our study with CT scans showing arterial extravasation was higher than anticipated. This finding likely reflects the improved diagnostic capability of helical CT. Although the spleen and liver were the organs most commonly associated with arterial extravasation, radiologists should be aware that arterial extravasation may be associated with several other visceral injuries.
View details for Web of Science ID 000172927900003
View details for PubMedID 11756079