CT of Ovarian Cancer for Primary Treatment Planning: What the Surgeon Needs to Know—Radiology In Training

Abstract

A 60-year-old woman presented with intermittent abdominal pain, an elevated serum CA-125 level, and an abnormal CT examination and was ultimately diagnosed with advanced-stage high-grade serous ovarian cancer. Key tumor locations on CT scans that should be highlighted by the radiologist to guide treatment selection are discussed.

© RSNA, 2022

Summary

CT imaging and disease-specific structured reporting are essential to guide primary treatment selection in patients with advanced ovarian cancer.

Teaching Points

• ■ Patients with ovarian cancer often present with advanced disease, for which complete resection at debulking surgery is the goal.

• ■ Pretreatment CT facilitates decision-making for optimal timing of debulking surgery and aids in presurgical planning, including the need for involvement of other surgical specialties.

• ■ The radiologist should provide a disease-specific structured report that highlights difficult-to-resect disease locations.

• ■ Research is ongoing on the value of fluorodeoxyglucose PET/CT, fluorodeoxyglucose PET/MRI, diffusion-weighted MRI, or multiparametric MRI as a replacement for or as an addition to CT for ovarian cancer treatment planning.

Dr Fernandes is currently a nuclear oncology fellow in the Molecular Imaging and Therapy Service, Radiology Department, Memorial Sloan Kettering Cancer Center in New York, New York. She completed her radiology residency in the Universidade Federal do Rio de Janeiro, Brazil, and abdominal imaging fellowship at the Centro de Imagem Rede D’Or, Rio de Janeiro, Brazil, followed by a body oncologic imaging fellowship at Memorial Sloan Kettering.

Dr Lakhman is an associate attending radiologist in the Body Imaging Service, Radiology Department, Memorial Sloan Kettering Cancer Center. She is an active member of the gynecologic oncologic disease management team and a director of gynecologic imaging. Dr Lakhman is enthusiastic about educating residents and fellows and participating in collaborative research.

Case Presentation (Dr Fernandes)

A 60-year-old woman first presented to a local physician with a month of intermittent abdominal pain. Her medical history was significant for chronic abdominal and back pain. Contrast-enhanced CT of the abdomen and pelvis was performed, demonstrating multiple mildly thickened nondilated fluid-filled loops of the small bowel. The patient was initially diagnosed with enteritis based on contrast-enhanced CT findings and prescribed prednisone for presumed Crohn disease. A month later, her symptoms worsened, including new onset of diarrhea and unintentional weight loss. She underwent MR enterography, which demonstrated large-volume ascites, thickened peritoneal reflections, and peritoneal nodules. Ascitic fluid cytology was positive for malignancy, favoring a primary müllerian tumor. The serum CA-125 level was elevated to 1928 U/mL (normal range: 0–35 U/mL), whereas both carcinoembryonic antigen and CA 19-9 levels were within normal limits at 1.4 U/mL (normal range: 0–5 U/mL) and 14 U/mL (normal range: 0–37 U/mL), respectively. At the gynecologic examination, there were palpable nodules in the cul-de-sac. The patient was referred to our institution for further management of suspected advanced ovarian cancer (OC).

CT Examination

Repeat contrast-enhanced CT scans demonstrated small-volume ascites and multifocal peritoneal carcinomatosis, including large omental cake (Fig 1). In the right upper quadrant, perihepatic implants were present in the right subdiaphragmatic and right posterior perihepatic regions, hepatorenal recess, falciform ligament, and fissure for ligamentum venosum. No hepatic parenchymal invasion or parenchymal hepatic metastases were seen. In the left upper quadrant, perisplenic implants were seen along the splenic capsule and ligaments, with no splenic hilar involvement or parenchymal metastases. In the pelvis, plaque-like peritoneal thickening was present and involved the cul-de-sac. No adnexal masses were identified. Small and large bowel mesentery were involved by tumor, including the central small bowel mesentery (mesenteric root), as indicated by tethering and nodules. Small and large bowel serosal implants were present in multiple locations. No abnormal findings were seen in the chest. The conclusion, following review of the CT scans, was that complete cytoreduction was not feasible due to disease in the central small bowel mesentery.

Figure 1:

Contrast-enhanced CT images demonstrate multisite peritoneal carcinomatosis. (A) Axial image shows implant in the fissure for ligamentum venosum (arrow). (B) Axial image demonstrates implants in the falciform ligament (arrow), hepatorenal recess (black arrowhead), and lesser sac (white arrowhead). (C) Axial image illustrates implants in the central mesentery (arrow). (D) Axial image shows implants in the small bowel mesentery (arrow) and bowel serosa (★). (E) Coronal image demonstrates implants in the right subdiaphragmatic region (black arrow), large bowel serosa (white arrows), and sigmoid mesentery (★). (F) Coronal image shows implants in the central small bowel mesentery (arrows) and large bowel serosa (stars). No abnormal findings were seen in the chest.

Diagnostic Laparoscopy

Diagnostic laparoscopy revealed right subdiaphragmatic implants and peritoneal carcinomatosis, including a 20-cm omental cake infiltrating the adjacent transverse colon. The small bowel was tethered and nonmobile, suggesting involvement of the central small bowel mesentery. Dense adhesions prevented evaluation of pelvic organs.

The surgical team concluded that upfront debulking surgery was not possible due to likely disease in the mesenteric root, with secondary high risk of mesenteric vascular injury and small bowel ischemia. Instead, the patient was referred for neoadjuvant chemotherapy (NACT) to reduce tumor burden. Biopsy samples taken during diagnostic laparoscopy revealed high-grade serous carcinoma.

Primary Treatment

Interval debulking surgery (IDS) was performed following three cycles of platinum-based NACT. Visual inspection at the start of laparotomy demonstrated excellent treatment response. Complete gross resection (CGR) was achieved after pelvic peritonectomy en bloc with uterus, adnexa, and appendix and resection of adhesions in the small bowel mesentery. After undergoing IDS, the patient completed three more cycles of platinum-based chemotherapy and remains disease-free 3 years from the initial treatment.

Case Discussion (Dr Lakhman)

What Is the Most Common Type of OC, What Are the Risk Factors for OC Development, and What Is the Pathogenesis?

Epithelial ovarian, fallopian, and primary peritoneal malignancies are categorized as a single entity due to their similar manifestation, staging, and management. Current evidence suggests common origin in the fimbriated end of the fallopian tube (1,2). Epithelial OC accounts for 90% of OCs, with serous, endometrioid, clear cell, and mucinous tumors as the main histologic subtypes (3,4). Of these, high-grade serous OC is the most common, observed in two-thirds of all OC cases, and is the leading cause of death from gynecologic cancer because of early and rapid peritoneal dissemination resulting in advanced disease at presentation (3,4).

Bloating, abdominal and pelvic pain, poor appetite, early satiety, and urinary symptoms are common at presentation. However, symptoms are often absent until disease is widespread (5). Risk factors for developing OC include family history of breast, ovarian, or colon cancer; genetic mutations like BRCA1 or BRCA2; nulliparity; endometriosis (associated with higher risk of endometroid and clear cell OC); postmenopausal status; and older age (6). The serum CA-125 level is routinely measured during the initial workup of clinically suspected OC. While it is elevated in 85% of patients with advanced OC, it is nonspecific and may also be elevated in nongynecologic malignancies and benign conditions (eg, endometriosis, pelvic inflammatory disease, and ovarian cysts) (7). Serum carcinoembryonic antigen and CA 19-9 levels may be obtained to help in the differentiation of primary mucinous OC from gastrointestinal tumor with ovarian or peritoneal spread (7). A CA-125–to–carcinoembryonic antigen ratio greater than 25 favors primary OC, while a CA-125–to–carcinoembryonic antigen ratio less than or equal to 25 suggests the need for a gastrointestinal workup including colonoscopy and/or upper endoscopy (8,9).

OC is staged according to the International Federation of Gynecology and Obstetrics classification system (Fig 2) (10). With no effective screening strategy and lacking early-stage symptoms, more than 80% of patients have ascites and extraovarian spread at presentation. OC spread follows three major pathways: peritoneal seeding, extension via lymphatics, and, less often, hematogenous dissemination. In advanced disease, the tumor is no longer confined to the ovaries (stage I) or true pelvis (stage II) but has disseminated to the abdominal peritoneum and/or retroperitoneal (pelvic and/or para-aortic) lymph nodes (LNs) (stage III), or spread to extra-abdominal LNs (supradiaphragmatic, inguinal, thoracic) and/or distant organs (stage IV) (eg, pleura, hepatic or splenic parenchyma) (10). Approximately two-thirds of all epithelial OCs manifest as stage III or IV disease (11).

Figure 2:

Illustration shows the International Federation of Gynecology and Obstetrics staging system for ovarian cancer. LN = lymph node.

What Are the Two Main Approaches for Primary Management of OC?

Traditional management consists of surgery and platinum-based chemotherapy. If disease is clinically confined to the adnexa, surgical goals include removal of ovaries and fallopian tubes (bilateral salpingo-oophorectomy), uterus and cervix (total abdominal hysterectomy), omentum (omentectomy), and retroperitoneal LNs, as well as peritoneal assessment using biopsies and cytology. In the more common scenario of advanced extraovarian disease, the surgical goal is complete removal of all visible and palpable disease, termed primary debulking surgery (PDS). Minimally, PDS includes total abdominal hysterectomy, bilateral salpingo-oophorectomy, and omentectomy, but may also require lymphadenectomy to remove enlarged LNs, upper abdominal surgery, and bowel resection (12). The size of the largest remaining lesion after surgery (residual disease) determines the outcome of PDS. Absence of visible or palpable residual disease indicates CGR; residual disease less than or equal to 1 cm signifies optimal resection; and residual disease greater than 1 cm indicates suboptimal resection. The aim of PDS is CGR because the two most important prognostic factors in advanced OC are residual disease status and response to chemotherapy (13–15). The goal in treatment planning is to avoid “futile” laparotomy resulting in suboptimal resection, as it incurs morbidity without survival gain. Because the CGR rate is proportional to surgical expertise, ideally a gynecologic oncologist with a high-volume practice should perform the surgery (12,16).

NACT followed by IDS (herafter, NACT-IDS) is an alternative to PDS. NACT could potentially decrease the extent of disease, reduce the complexity and morbidity of debulking surgery, and increase the likelihood of CGR. The optimal timing of surgery was investigated in several clinical trials where patients with advanced OC were randomized to either PDS or NACT-IDS (17–21). These trials found a higher rate of CGR and lower risk of complications in the NACT-IDS group, while progression-free survival and overall survival were similar between the PDS and NACT-IDS groups. As concerns have been raised regarding substandard surgical techniques in these trials, the Trial of Radical Upfront Surgery (TRUST), an ongoing prospective, international multicenter trial emphasizing surgical quality, will clarify the optimal timing of surgery (22). Meanwhile, the National Comprehensive Cancer Network practice guidelines and European Society of Gynecological Oncology guidelines for OC surgery recommend PDS if the patient is fit for surgery and CGR (or at least optimal resection) is feasible (12,23). The guidelines advise reserving NACT-IDS for patients who are poor surgical candidates or if the disease burden likely prohibits CGR (or at least optimal resection). This strategy highlights the importance of proper patient selection.

How Is CT Performed and Used to Triage Patients with Advanced OC to Appropriate Primary Treatment?

Multiple studies have explored the role of imaging-based tools alone or in combination with clinical factors to predict resectability, arriving at mixed conclusions (24–36). The potential explanation is that beyond disease burden (reflected on imaging) and performance status, resectability is influenced by surgical expertise, institutional philosophy about acceptable surgical morbidity, and patient preference. These factors, and the lack of universal resectability criteria, result in variable CGR feasibility across centers. Presently, the best approach to primary treatment selection (PDS vs NACT-IDS) is a comprehensive review of clinical and imaging data followed by a multidisciplinary discussion. Radiologists play an integral role in this decision-making process because their reports highlight tumor locations that may present surgical challenges, increase the risk of complications, or require intraoperative input and support from other surgical consultants. If, after multidisciplinary review, the feasibility of CGR remains uncertain, diagnostic laparoscopy is considered (Fig 3).

Figure 3:

Algorithm for triaging patients with suspected advanced ovarian cancer to primary treatment (ie, primary debulking surgery vs neoadjuvant chemotherapy followed by interval debulking surgery). CEA = carcinoembryonic antigen, iv = intravenous, OC = ovarian cancer.

CT of the chest, abdomen, and pelvis is the mainstay of imaging to stage advanced OC. It should be performed with oral and intravenous contrast materials unless contraindicated (12). Positive oral contrast material may help to visualize bowel loops and detect peritoneal deposits, in particular serosal and mesenteric implants (37,38). An important caveat is that small, calcified implants (more common with low-grade serous OC) may be obscured by positive oral contrast material and are more readily identified with negative oral contrast material, such as water (38).

The accuracy of CT for facilitating detection of peritoneal dissemination is high; however, its sensitivity varies across anatomic sites and between readers (39–44). For example, a recent study reported that sensitivity for omental dissemination, subdiaphragmatic implants, and bowel involvement (serosa, mesentery) ranged from high (90%–98%) to moderate (43%–93%) to low (38%–55%), respectively (43). CT sensitivity is reduced further when the tumor is small (40,43). Review of reconstructed coronal and sagittal images in addition to standard axial images may aid detection. CT has moderate sensitivity (43%), albeit high specificity (95%), as a tool for detection of LN metastases, relying primarily on the size criterion of greater than 1 cm in the short-axis dimension to diagnose LN enlargement (45,46).

Staging CT reports should describe the presence and volume of ascites and the presence, location, and extent of adnexal masses, peritoneal implants, bowel involvement (serosa, mesentery), LN enlargement, pleural effusions, pleural nodules, and distant metastases. The size of the largest lesion in each location should be reported and a disease-specific structured report is recommended. From the surgical perspective, some disease sites are more challenging to clear than others; these potentially difficult-to-resect tumor locations deserve particular mention, as emphasized hereafter and highlighted in Figure 4. Disease-specific reporting and standard lexicon can help radiologists include all key disease locations in a clear and organized manner (47,48). In fact, gynecologic surgical oncologists find disease-specific reports easier to understand and more useful for presurgical planning compared with even simple structured reports (48).

Figure 4:

Chart shows disease locations and likelihood of resectability. No universally accepted resectability criteria exist and, thus, there is significant variability across centers.

Which Features on CT Scans Should Be Flagged by the Radiologist for the Surgical Oncologist?

When interpreting CT scans, a radiologist should emphasize disease locations that may pose a challenge to complete surgical resection, increase the risk of complications, or require input from other surgical consultants (Fig 4).

Peritoneal dissemination on CT scans can appear as thickened peritoneal reflections, subtle peritoneal fat infiltration, or peritoneal nodules and/or masses (Fig 1). The presence of ascites should raise suspicion for peritoneal disease (49).

In the right upper quadrant, perihepatic implants located along the subdiaphragmatic region, right and left hepatic surfaces, falciform ligament, hepatorenal recess (Morison pouch), and gallbladder fossa can usually be resected, particularly if a surgeon is aware of the disease in these locations and the potential need for liver mobilization. Implants in the fissure for ligamentum venosum (between the caudate lobe and left hepatic lobe), porta hepatis, lesser omentum (hepatogastric and hepatoduodenal ligaments), and perihepatic implants with parenchymal invasion may require more extensive hepatic resection and/or assistance from a hepatobiliary surgeon (50) (Figs 4–6). Parenchymal invasion is suggested when a perihepatic lesion protrudes into the hepatic parenchyma and demonstrates an ill-defined, nodular, or obliterated implant-liver interface (51). Lesser sac implants may be a challenge to clear because their presence signifies widespread disease and potential invasion of adjacent organs, such as the stomach or pancreas (30).

Figure 6:

Diagram illustrates the lesser sac and its borders. The lesser sac is bordered anteriorly by the liver, stomach, and greater omentum and posteriorly by the pancreas, left adrenal gland, and left kidney; on the left side by the splenorenal and gastrosplenic ligaments and on the right side by the lesser omentum. The lesser omentum joins the stomach and proximal duodenum to the liver, includes hepatogastric and hepatoduodenal ligaments, and encloses the porta hepatis. Splenic ligaments include the splenorenal, gastrosplenic, and splenocolic ligaments (latter not illustrated). Tumor implants in the perihepatic region, porta hepatis, lesser sac, and splenic hilum are shown in purple.

Figure 5:

Diagram illustrates the peritoneal spaces. The peritoneal cavity is divided into the greater peritoneal cavity and lesser sac (omental bursa). The greater omentum is a multilayered fold of peritoneum that extends down from the greater curvature of the stomach. The gastrocolic ligament is a portion of greater omentum between the stomach and transverse colon. The lesser omentum is a multilayered fold of peritoneum that extends from the lesser curvature of the stomach and proximal duodenum to the liver. A pelvic peritoneal recess between the uterus and rectum (rectouterine recess) is known as the cul-de-sac or pouch of Douglas.

In the left upper quadrant, implants in the left subdiaphragmatic region, along the splenic capsule and ligaments (gastrosplenic, splenorenal, splenocolic), can often be resected, although the surgeon should be alerted in advance to the potential need for splenic mobilization. Involvement of the splenic hilum or parenchyma indicates the need for splenectomy, which if anticipated, can be performed en bloc with the omentum (Figs 4, 6). If the need for splenectomy is anticipated based on CT findings, the patient can undergo all necessary vaccinations before debulking surgery to reduce the risk of a life-threatening postsurgical infection.

Involvement of the bowel serosa and/or mesentery may present a major barrier to CGR, as there is a physiologic limit to how much bowel can be removed without causing unacceptable morbidity (23). Identifying mesenteric involvement or serosal implants can be a challenge because subtle infiltration of mesenteric fat, bowel angulation and/or retraction, or small nodules can be overlooked. While the rectosigmoid colon is often involved by contiguous tumor extension from the cul-de-sac or adnexa, this can be addressed easily with a low anterior resection en bloc with total abdominal hysterectomy and bilateral salpingo-oophorectomy. Limited or peripheral involvement of the mesentery and limited involvement of the bowel, stomach, or duodenum can be managed with short segment bowel resection or limited gastric or duodenal resection, respectively (Figs 4, 7). However, more extensive disease, including infiltration of the central small bowel mesentery, tumor requiring long segments of small bowel resection, involvement of more than two segments of colon necessitating total colectomy, or total gastrectomy, precludes CGR due to unacceptable morbidity (Figs 4, 7) (23). Uncommon involvement of the celiac artery, hepatic artery, and left gastric artery is also considered unresectable (23).

Figure 7:

Diagram illustrates the relationship between the extent of small bowel mesenteric and/or serosal involvement and feasibility of complete cytoreduction. The small bowel mesentery is a large fat-laden peritoneal reflection that attaches the jejunum and ileum to the posterior abdominal wall via the root of small bowel mesentery. The root runs on a diagonal from the duodenojejunal junction to the ileocecal region and contains superior mesenteric vessels, nerves, and lymphatics.

While all abnormal LNs should be mentioned in the radiologic report (ie, >1 cm in the short-axis dimension or with suspicious morphologic features such as a round shape or necrotic component), some LN stations present greater surgical challenges, including LNs in the para-aortic region above the renal vessels, in the porta hepatis, next to the celiac artery, in the mesentery, and in the retrocrural region (Fig 4).

Supradiaphragmatic (cardiophrenic) LNs receive lymphatic drainage from the abdomen and, thus, are a common site of LN metastases in advanced OC (Fig 8). They are considered abnormal if greater than 0.5 cm in the short-axis dimension (52). Despite their location outside the abdominal cavity (ie, stage IVB), supradiaphragmatic LNs can be resected by using the transdiaphragmatic approach, if anterior. Similarly, inguinal LNs greater than 1.5 cm in the short-axis dimension are considered abnormal but can also be resected despite their stage IVB designation (Fig 4) (46).

Figure 8:

Diagram illustrates the supradiaphragmatic and retrocrural lymph node groups. Supradiaphragmatic lymph nodes are divided into two major groups: anterior and middle. The anterior group is located posterior to the xiphoid process and just behind the anterior seventh rib costochondral junction. The middle group is usually present on the right side and absent on the left side. Retrocrural lymph nodes are situated posterior to each crus of the diaphragm.

In addition to supradiaphragmatic and inguinal LNs, several other sites of stage IV disease can be resected, such as splenic parenchymal metastases with splenectomy. In contrast, parenchymal hepatic, pulmonary, pleural, and abdominal wall and/or soft-tissue metastases are generally considered unresectable unless they are limited or solitary in nature (23). Likewise, supraclavicular, mediastinal, and retrocrural LN metastases usually indicate that CGR is not feasible (Fig 4).

What Is the Role of MRI and PET/CT in Primary Treatment Selection for Patients with Advanced OC?

CT is the current standard-of-care imaging technique used to assess disease extent and evaluate the feasibility of PDS because it is widely available and fast. Whole-body fluorodeoxyglucose (FDG) PET and MRI are two additional imaging techniques that are increasingly available in many centers. Their role in pretreatment evaluation of patients with OC is under investigation.

Conventional MRI (ie, T1-weighted, T2-weighted, and contrast-enhanced sequences) has superior soft-tissue contrast compared with CT. Diffusion-weighted imaging (DWI), if combined with conventional MRI (hereafter, DWI MRI), further improves soft-tissue contrast. Peritoneal implants, LNs (benign or malignant), and metastases demonstrate high signal intensity on high-b-value DWI scans, resulting in increased conspicuity. Compared with CT, DWI MRI improves the assessment of peritoneal tumor burden and resectability (53–57). For example, a single-center prospective study of 94 patients with OC found that DWI MRI was more accurate compared with CT for staging (87% DWI MRI vs 35% CT) and predicting incomplete resection at PDS (96% DWI MRI vs 71% CT) (54). DWI MRI has also been shown to be superior to CT or FDG PET in the challenging evaluation of mesenteric and serosal disease (53,54,58). Nevertheless, MRI is less widely available, requires a longer scan time, and can be susceptible to artifacts caused by patient motion, peristalsis, cardiac motion, and bowel gas; thus, it demands greater expertise for high-quality implementation and interpretation (59,60).

FDG PET/CT is more accurate than CT or MRI in the detection of LN metastases and extra-abdominal spread (45,61,62). This may be particularly helpful in evaluating LNs that are not enlarged according to size criteria. A meta-analysis demonstrated a sensitivity and specificity of 73% and 96%, respectively, for FDG PET/CT compared with 42% and 95% for CT and 54% and 88% for MRI (45). Nevertheless, FDG PET/CT is not superior to CT or MRI for assessing peritoneal dissemination. FDG PET/CT is limited for detecting subcentimeter peritoneal implants (especially <5 mm) due to low spatial resolution (63), identifying mucinous implants because of their absent-to-low FDG uptake (64), and diagnosing mesenteric and serosal implants secondary to physiologic tracer excretion in the bowel and urine (65,66).

There is emerging technology allowing PET to be obtained simultaneously with and fused to MRI, with potential advantages of combined anatomic information from conventional MRI, functional information from DWI MRI and dynamic contrast-enhanced MRI, and metabolic data from PET. To date, most studies in gynecologic oncology have focused on the role of PET/MRI in the recurrent rather than initial setting. A recent study of 34 patients with peritoneal carcinomatosis from ovarian or endometrial cancer found that PET/MRI was superior to DWI MRI alone in estimating peritoneal disease extent, especially in patients with high tumor burden (67).

Lastly, a recent meta-analysis on the role of FDG PET/CT, conventional MRI, and DWI MRI as a replacement for or as an addition to CT for assessment of tumor extent and resectability concluded that, at present, there is insufficient evidence (due to small sample size and variable methodology) to advise the routine addition of these techniques to clinical practice (68). At this time, the National Comprehensive Cancer Network recommends MRI and FDG PET (fused to CT or MRI) for characterizing indeterminate lesions if the results will alter management (12). In contrast, the joint European Society of Medical Oncology and European Society of Gynecological Oncology guidance is that CT, FDG PET/CT, or whole-body DWI MRI can be used as part of the pretreatment workup to determine disease extent (69).

Looking ahead, the MR in Ovarian Cancer study (ISRCTN51246892), a large ongoing, prospective, and multicenter clinical trial of patients with clinically suspected or confirmed OC, may clarify the added value of multiparametric MRI (ie, MRI with conventional sequences, DWI MRI, and dynamic contrast-enhanced MRI) for the diagnosis, staging, and resectability evaluation of OC (70).

In conclusion, most patients with ovarian cancer present with advanced stage disease, for which complete resection at debulking surgery is the goal. CT plays an important role in assessing disease extent and selecting a primary treatment approach (ie, primary debulking surgery vs neoadjuvant chemotherapy plus interval debulking surgery). The radiologic report can serve as a roadmap to facilitate patient triage and clinical decision-making by alerting gynecologic oncologists to the key disease locations that may complicate or preclude complete (or at least optimal) debulking surgery. Research is ongoing to determine the value of other imaging modalities as a replacement for or as an addition to CT.