25. Pelvic CT

Urologic CT interpretation is governed by a single principle: the urinary tract is a contrast-conducting system whose pathology declares itself differently in each contrast phase, so the diagnostic question dictates the protocol. The reference examination is CT urography (CTU), a multiphase study in which an unenhanced acquisition interrogates calculi and baseline attenuation, a nephrographic phase ($90$–$100\,\text{s}$) maximally opacifies the renal parenchyma to detect masses, and an excretory (urographic) phase ($8$–$15\,\text{min}$) fills the collecting systems, ureters, and bladder to map urothelial disease. The ACR Appropriateness Criteria and the AUA/SUFU microhematuria guideline operationalize when this is justified: gross hematuria and high-risk microhematuria warrant CTU, whereas low-risk microhematuria may be served by ultrasound, reflecting an explicit pretest-probability calculus for genitourinary malignancy.

Urolithiasis is the highest-volume urologic diagnosis and the cleanest example of mechanistic CT reading. A stone is a crystalline concretion that obstructs urine flow, and noncontrast CT is essentially $100\%$ sensitive because virtually all stones — even radiolucent ones on plain film — exceed soft-tissue attenuation. Composition is encoded in density and behavior: calcium oxalate/phosphate stones are very dense (often $>600$–$1000\,\text{HU}$), struvite (infection) stones are softer and form staghorn casts of the collecting system, cystine stones are moderately dense ($\sim 400$–$600\,\text{HU}$) and faintly visible on film, and uric-acid stones ($\sim 200$–$450\,\text{HU}$) are the lone radiolucent group that may be dissolved by urinary alkalinization. Dual-energy CT resolves this prospectively by material decomposition, separating uric-acid from calcium stones by their differential attenuation at two photon energies and directly redirecting therapy. The secondary signs of obstruction — hydroureteronephrosis, perinephric and periureteric fat stranding, and the soft-tissue rim sign of an edematous ureteral wall encircling an impacted calculus — confirm that a calcification lies within the lumen rather than a phlebolith beside it; phleboliths characteristically show a central lucency and a comet-tail sign of an adjacent vessel. Management is governed by size and site: stones $\le 5\,\text{mm}$ pass spontaneously in the large majority, $5$–$10\,\text{mm}$ stones are intermediate, and stones $>10\,\text{mm}$ or those at the ureteropelvic or ureterovesical junctions are far less likely to pass and prompt urologic intervention.

Upper-tract malignancy is the reason the excretory phase exists. Urothelial (transitional cell) carcinoma of the renal pelvis or ureter appears as an enhancing soft-tissue filling defect, focal urothelial thickening, or a stricture, and its hallmark is enhancement that distinguishes it from a non-enhancing clot or stone. Renal cell carcinoma, by contrast, is a parenchymal mass; the clear-cell subtype enhances avidly and heterogeneously (an absolute attenuation increase well beyond $20\,\text{HU}$), and the search pattern must extend to the renal vein and IVC for tumor thrombus and to the retroperitoneal nodes for staging. Bladder cancer presents as focal wall thickening or a polypoid enhancing mass; CT stages nodal and distant disease and detects perivesical fat invasion, but it cannot reliably resolve the critical T1–T2 question of detrusor-muscle invasion, which is the province of multiparametric MRI scored by VI-RADS and of cystoscopic resection. The dominant failure modes are characteristically urologic: a renal cyst that pseudoenhances by $10$–$20\,\text{HU}$ from beam hardening can be mistaken for a hypovascular tumor; partial-volume averaging mimics urothelial thickening; layering of dense excreted contrast can bury a small bladder tumor (mandating prone or delayed views); and the cognitive trap of satisfaction of search lets an obvious obstructing stone halt the read before the contralateral small renal mass is found.

Although MRI and ultrasound are the dedicated tools of the female pelvis, CT is overwhelmingly the index study in the acute abdomen and in oncologic staging, so the interpreter must extract maximal gynecologic information from a contrast-enhanced volume and know precisely when to defer. The interpretive frame is anatomic and physiologic: the premenopausal uterus is a vascular, avidly enhancing organ whose junctional zone is invisible on CT, the ovaries are follicle-bearing structures whose physiologic cysts are normal, and the great diagnostic errors arise when normal cyclical change is read as disease or when a surgical emergency is mistaken for a benign cyst.

The acute female pelvis is fundamentally a Bayesian problem. Ovarian torsion is a mechanical strangulation: the ovary twists on its vascular pedicle, venous and lymphatic outflow fail first, and the CT signature is an enlarged, edematous ovary (often $>4\,\text{cm}$) deviated toward the midline, with peripherally displaced follicles, a twisted vascular pedicle (the whirlpool sign), and — as arterial inflow finally fails — absent or diminished enhancement and adjacent free fluid. Because torsion is a clock-dependent surgical emergency, even a moderate posterior probability mandates immediate gynecologic consultation. Tubo-ovarian abscess, the end-stage of ascending pelvic inflammatory disease, appears as a thick-walled, rim-enhancing complex adnexal collection, frequently with internal gas, surrounding fat stranding, and inflammatory thickening of the fallopian tube; its differential against an endometrioma or a torsed cyst is resolved by the clinical triad of fever, leukocytosis, and cervical-motion tenderness, which sharply raises the pretest probability of infection. A ruptured ovarian (hemorrhagic) cyst produces pelvic free fluid that may be hyperattenuating ($>30\,\text{HU}$) when frankly hemorrhagic, with a collapsing cyst and, in significant bleeds, a sentinel clot of highest attenuation adjacent to the source. Ectopic pregnancy is a clinical and β-hCG diagnosis confirmed by ultrasound, but CT performed for undifferentiated pain may reveal an adnexal mass with hemoperitoneum, and the unwary reader who anchors on a non-gynecologic cause courts catastrophe. The essential extragynecologic mimics — appendicitis and sigmoid diverticulitis — are separated by tracing the inflamed structure to its origin, the single most reliable discriminator in the inflamed pelvis.

Gynecologic malignancy on CT is principally a staging exercise grounded in defined systems. Endometrial carcinoma manifests as a thickened, heterogeneously hypoenhancing endometrium, often with fluid distending an obstructed cavity, but the depth of myometrial invasion that drives FIGO staging is an MRI question. Cervical carcinoma is staged by the FIGO 2018 system, which — in a major revision — incorporates nodal status (stage IIIC) and explicitly permits imaging and pathology to assign stage; CT's strengths are nodal and distant disease and the detection of hydronephrosis (which alone defines stage IIIB), while parametrial invasion is better resolved by MRI. Ovarian carcinoma typically presents late as a complex cystic-solid adnexal mass with peritoneal carcinomatosis, omental caking, and ascites, and CT is the workhorse for mapping disease burden to guide primary debulking versus neoadjuvant chemotherapy. Risk stratification of the adnexal mass itself is governed by O-RADS and the IOTA ADNEX model, whose morphologic logic — solid enhancing components, thick irregular septa, and papillary projections raise malignant probability — transfers usefully to CT morphology even though those systems were derived for ultrasound and MRI. The failure modes are again perceptual and cognitive: a physiologic corpus-luteum cyst with a crenulated enhancing wall mimics a neoplasm or abscess; a pedunculated uterine leiomyoma simulates an adnexal mass until its bridging vessels to the uterus are traced; dependent bowel loops simulate adnexal lesions; and premature closure on a “cyst” label is the recurring error that allows torsion and carcinoma to be discharged.

CT is a structurally blind instrument for the internal architecture of the prostate. Because the gland's zonal anatomy — the peripheral zone where the majority of carcinomas arise, the central zone, and the transition zone that hypertrophies with age — is essentially isoattenuating to soft tissue on CT, the modality cannot reliably detect, localize, or locally stage intraprostatic cancer; that role belongs unequivocally to multiparametric MRI interpreted with PI-RADS v2.1, the governing framework that scores diffusion restriction, dynamic enhancement, and T2 morphology to assign the probability of clinically significant disease. The disciplined CT reader therefore frames prostate interpretation around three questions CT can answer: gross glandular enlargement and its downstream effects, infection and abscess, and the locally advanced or metastatic spread of established malignancy.

Benign prostatic hyperplasia is a histologic proliferation of the transition zone that on CT presents simply as a symmetrically enlarged gland, sometimes with a median lobe indenting and elevating the bladder base, calcifications along the surgical capsule, and the secondary consequences of bladder-outlet obstruction — detrusor wall thickening and trabeculation, diverticula, and in advanced cases bilateral hydroureteronephrosis. The mechanistic chain from obstruction to upper-tract dilatation is the clinically important read, because it converts an incidental enlarged prostate into an actionable finding. Acute bacterial prostatitis and prostatic abscess — the latter most common in diabetic or immunocompromised men and after instrumentation — appear as glandular enlargement with heterogeneous or frankly low-attenuation, rim-enhancing fluid collections; an organized abscess ($\sim 0$–$25\,\text{HU}$ center with an enhancing wall) is a drainable lesion, and recognizing it on a CT obtained for sepsis of unknown source can be decisive. Emphysematous prostatitis, with intraglandular gas, is a surgical emergency analogous to its renal counterpart.

Prostate carcinoma is encountered on CT not at the moment of detection but at the moment of staging and surveillance. Locally advanced disease may efface the rectoprostatic (Denonvilliers) fat plane, asymmetrically enlarge or obliterate a seminal vesicle (loss of the normal seminal-vesicle angle), or invade the bladder base. CT's genuine strength is nodal and skeletal staging: pathologic pelvic and retroperitoneal adenopathy follows a predictable obturator-and-iliac chain distribution, and prostate cancer's profound osteotropism produces the most characteristic of all CT signatures — osteoblastic (sclerotic) metastases to the pelvis, lumbar spine, and proximal femora, dense bone-forming deposits that are nearly pathognomonic in an older man with a known or suspected prostate primary. This osteoblastic preference is itself mechanistic, driven by tumor-secreted factors that activate osteoblasts, and it stands in deliberate contrast to the lytic deposits of renal and thyroid carcinoma. Therapeutic response is tracked with RECIST 1.1, with the important caveat that purely osteoblastic bone disease is considered non-measurable, a recurring source of response-assessment confusion. The dominant failure modes are the predictable consequences of CT's contrast blindness and pelvic physics: the false reassurance of a normal-appearing gland (CT cannot exclude organ-confined cancer and must never be used to do so); streak and beam-hardening artifact from a hip prosthesis or dense rectal contrast that obscures the gland and seminal vesicles; mistaking a distended seminal vesicle or a Cowper-gland structure for adenopathy; and the anchoring error of attributing all sclerotic foci to benign bone islands when an osteoblastic metastatic pattern is emerging.

The indeterminate pelvic mass is the synthetic endpoint of this chapter, because the radiologist's first and most consequential task is not naming the lesion but assigning it to an organ — the organ-of-origin determination that collapses an unmanageable differential into a tractable one. The governing maneuvers are mechanical and reproducible: the beak sign (the parent organ's tissue forming an acute angle that claws around the mass) and the embedded-organ and phantom-organ signs localize origin; identifying the feeding vascular pedicle (as in a pedunculated leiomyoma traced to the uterus, or the ovarian vessels traced to an adnexal mass) confirms it; and the displacement pattern of adjacent structures — bowel, bladder, and ureters — indicates whether a mass arises from the pelvic viscera, the peritoneum, or the retroperitoneum. Only after origin is fixed does tissue characterization by attenuation become diagnostic.

Attenuation is the second axis of analysis, and a handful of values are near-specific. Macroscopic fat ($<-30$ to $-100\,\text{HU}$) within a mass is the signature of a benign mature cystic teratoma (dermoid) when combined with calcium/tooth elements and a fat–fluid level, or of a benign lipomatous lesion; its recognition essentially closes the differential. Simple fluid attenuation ($\sim 0$–$20\,\text{HU}$) in a thin-walled non-enhancing lesion characterizes a benign cyst, whether ovarian, a peritoneal inclusion cyst conforming to its space, or a lymphocele in the postoperative pelvis. Hyperattenuating fluid ($>30\,\text{HU}$) suggests hemorrhage (an endometrioma or hemorrhagic cyst) or, when loculated with an enhancing wall and gas, an abscess. Avid heterogeneous enhancement with solid components, thick irregular septa, and papillary projections raises the posterior probability of malignancy, the morphologic logic formalized by O-RADS and the IOTA ADNEX model for adnexal lesions and broadly transferable to CT.

The ranked differential for a pelvic mass is therefore disciplined by Bayesian priors that depend on sex, age, and compartment. In a reproductive-age woman, gynecologic causes dominate: a uterine leiomyoma (the commonest solid pelvic mass, often calcified or showing cystic degeneration), a physiologic or hemorrhagic ovarian cyst, an endometrioma, or a dermoid. In an older woman, an enlarging complex cystic-solid adnexal mass shifts the prior decisively toward ovarian carcinoma, the cardinal CT signatures of which extend beyond the primary to peritoneal carcinomatosis — omental caking, scalloping of the liver and splenic surfaces by implants, and ascites — a pattern shared with gastrointestinal primaries and mesothelioma and itself a staging diagnosis assessed for resectability and tracked by RECIST 1.1. Compartment-specific entities sharpen the list: midline cystic lesions raise urachal remnants and anterior meningoceles; presacral masses in adults evoke chordoma, the lytic/destructive sacral tumor, schwannoma, and — in the appropriate setting — metastatic adenopathy; and bulky low-attenuation or necrotic nodal masses suggest lymphoma or metastatic disease. The failure modes in pelvic-mass interpretation are the most instructive in the chapter. Pseudoenhancement and partial-volume averaging falsely solidify a benign cyst; a pedunculated leiomyoma, an exophytic renal mass, and a bowel-origin GIST are each routinely misassigned when origin is not deliberately established; calcification is over-read as benignity when malignant psammomatous or osteoid calcification is the real substrate; and the dominant cognitive errors — anchoring on the largest mass, premature closure on a benign label, and satisfaction of search after the index lesion is found — are precisely the biases that allow carcinomatosis, nodal disease, and a second synchronous lesion to escape the report.