Computed tomography findings of peritoneal tumour in the absence of primary lesions: preliminary study

Abstract

Primary peritoneal tumours are uncommon and include mesothelioma and carcinoma. Mesothelioma identification is challenging, as there are no established rapid diagnostic methods. In the absence of a simple means of diagnosis, evaluation often requires several invasive biopsies, highlighting the need to develop a definitive diagnostic method. To address this, we focused on identifying characteristic computed tomography (CT) findings of primary peritoneal tumours. This study retrospectively assessed CT findings of peritoneal tumours in the absence of primary lesions elsewhere, including four carcinomas, five mesotheliomas, and three lymphomas in five dogs and seven cats. Whole‐body contrast‐enhanced CT was performed with data obtained on peritoneal tumour location, shape, contrast enhancement patterns during the arterial, portal and equilibrium phases, lymphadenopathy, ascites, calcification and attenuation values of peritoneal tumours among precontrast and postcontrast images. For carcinomas, 25% of cases were located in the visceral region and 75% in both parietal and visceral regions. All cases of mesothelioms were identified in the parietal region, while 80% were identified on the diaphragm. For Lymphomas, 67% of cases were located in the parietal region and 33% of cases in the visceral region. Location was a characteristic feature, showing significant differences among tumours (P = 0.006). Among tumours, no significant differences were observed for tumour shape, contrast enhancement patterns, lymphadenopathy, ascites and calcification. For mesothelioma, the attenuation value was significantly high in the portal and equilibrium phases images compared to precontrast images (P = 0.04). Thus, tumour location and attenuation value may assist in differentiating peritoneal tumours.

Abbreviations

CT

Computed Tomography

HU

Hounsfield Unit

US

Ultra Sonography

Peritoneal tumours are classified as either primary or secondary, ¹ with secondary peritoneal tumours including carcinomatosis, lymphomatosis and sarcomatosis. ² While secondary peritoneal tumour diagnosis is established and performed through primary lesion identification, ³ primary peritoneal tumours are uncommon ¹ with less certain means of diagnosis. These tumours include mesothelioma, ¹ which originates from mesothelial cells of the pleura, pericardium or peritoneum. ⁴ , ⁵ While pleural and pericardial occurrences are common, peritoneal occurrences account for 12.5%. ⁶ As there is no established method for mesothelioma diagnosis, identification is challenging. In the absence of a definitive diagnostic method, evaluation often requires several invasive biopsies. Conventional mesothelioma biopsy techniques include luminal fluid cytology by intraluminal puncture, computed tomography (CT)‐guided fine‐needle aspiration of suspicious lesions, surgery, thoracoscopy and laparoscopy. ⁴ , ⁷ , ⁸ However, these techniques have low sensitivity due to the limited tissue sampling available. ⁴ , ⁷ , ⁸ Diagnosing mesothelioma via luminal fluid cytology is also challenging, as distinguishing it from reactive mesothelial cells is difficult. ⁵ , ⁹ Additionally, as mesothelial cells resemble epithelial cells, differentiating mesothelioma and adenocarcinoma is further challenging, especially when both are present at metastatic sites. ¹⁰

In veterinary medicine, CT evaluation is useful for differentiation of tumour. ¹¹ , ¹² , ¹³ However, to the best of our knowledge, reports on CT findings of peritoneal mesothelioma are limited. For pleural mesothelioma in dogs and cats, many CT images fail to detect gross lesions, nodules or thickening lesions. ⁴ , ⁶ , ¹⁴ As primary peritoneal carcinoma occurrence is reported in cats, ¹⁵ we focused on identifying characteristic CT findings to assist in primary peritoneal tumour identification. This study retrospectively assessed CT findings of peritoneal tumours without primary lesions, including carcinoma, mesothelioma and lymphoma.

Materials and methods

This study employed retrospective cases. Owners of animals included in this study provided their informed consent for diagnostic procedures, treatment and use of clinical data, such as medical history, imaging studies and histopathological findings, for research and publication purposes. As diagnostic studies and initiated treatments were part of normal clinical activities, ethics and welfare committee evaluation was not required.

All animals with peritoneal tumours who had undergone CT at the Veterinary Medical Centre of Osaka Metropolitan University and the Kinki Animal Medical Training Institute between 2016 and 2025 were enrolled in the current study. The inclusion criteria were the presence of concurrent histopathologically diagnosed peritoneal lesions. Exclusion criteria required the presence of a primary lesion.

All animals underwent CT scanning with multidetector 16‐slice CT scanners (SOMATOM Scope; SIEMENS, Tokyo or Activion16; Canon Medical Systems Corporation, Tochigi, Japan) in helical scan mode. CT and reconstruction filter settings were based on our previous protocol. ¹² All animals underwent general anaesthesia, were ventilated and were placed in a supine position before undergoing CT. For all animals, breath‐hold was induced during image acquisition for the whole body scans, including head and neck region, chest, abdomen and the perineum. The injection time was 20 s. Contrast‐enhanced studies were performed in the arterial (20 s after contrast injection), portal (60 s after contrast injection) and equilibrium (180 s after contrast injection) phases.

All CT scans were reviewed by two experienced veterinary radiologists with CT features documented by consensus. These radiologists were not aware of the final diagnoses at the time of CT image review. Images were assessed in random order over three different readout sessions with a 2‐week minimum interval between each session to minimise potential bias.

Qualitative CT data were analysed and the following parameters noted: peritoneal tumour location (parietal, visceral or parietal and visceral), tumour shape (localised mass lesion, thickening), patterns of tumour contrast enhancement (homogeneous or heterogeneous) in the arterial, portal and equilibrium phases, presence or absence of lymphadenopathy, ascites and calcification. In accordance with literature, ¹⁶ the enhancement pattern was classified as homogeneous if the affected segment showed less than 10 Hounsfield units (HU) enhancement or heterogeneous if the enhancement was greater. Attenuation, in HU, was measured while avoiding blood vessels, cystic, necrosed areas and artefacts. The tumour enhancement pattern for the arterial, portal and equilibrium phases was similarly defined as homogeneous or heterogeneous. Lymphadenopathy was determined present when lymph node length or width was greater than the mean size of normal lymph nodes, ¹⁷ , ¹⁸ with the presence and lymphadenopathy location recorded.

Quantitative CT data were analysed in conjunction with the peritoneal tumour attenuation values. Tumour attenuation, in HU, was measured on precontrast, arterial, portal and equilibrium phases postcontrast images by manually drawing three ROIs to include the tumour but exclude cystic and necrosed areas. Measurements were repeated to evaluate the consistency and reliability of the values. The median HU in the precontrast, arterial, portal and equilibrium phase postcontrast images were calculated.

Statistical analyses

Statistical analyses were performed using commercially available software (R version 2.12.1; R Foundation for Statistical Computing, Vienna, Austria). Data normality was assessed using the Shapiro‐Wilk test, indicating that non‐parametric testing was required. Differences in qualitative parameters among peritoneal tumours were assessed using Fisher's exact test. Attenuation values of each peritoneal tumour among precontrast and postcontrast images were compared using the Friedman test and Steel‐Dwass post hoc test. A P‐value of <0.05 was considered statistically significant.

Result

Five dogs and seven cats were used in this study. Peritoneal tumours were diagnosed as carcinoma (n = 4, 33%), mesothelioma (n = 5, 42%) or lymphoma (n = 3, 25%). For carcinomas, one case (1 of 4; 25%) was diagnosed by fine needle aspiration (FNA), one (1 of 4; 25%) by Tru‐cut needle biopsy, and two (2 of 4; 50%) by cell block of luminal fluid. For mesotheliomas, one case (1 of 5; 20%) was diagnosed by Tru‐cut needle biopsy, and four (4 of 5; 80%) by cell block of luminal fluid. For lymphomas, one case (1 of 3; 33%) was diagnosed by cell block of luminal fluid, and two (2 of 3; 67%) by FNA. The carcinoma cases consisted of one intact male cat, one intact female cat and two spayed cats. One case presented with only abdominal distension, and three cases presented with anorexia. The median age of the cats in the carcinoma group was 11 years (range, 8–12 years). All the cats were of mixed breed. The mesothelioma cases consisted of one neutered male cat, one neutered male dog, one intact female dog and two spayed female dogs. Two cases presented with abdominal distension, lethargy and anorexia, and three cases presented with only abdominal distension. The median age of the dogs in the mesothelioma group was 8 years (range, 7–15 years). Dog breeds included one each of Spitz, German Shepherd, Mixed and Yorkshire Terrier. All the cats were of mixed breed. The lymphoma cases consisted of one neutered male dog and two neutered male cats. Two cases presented with only abdominal distension, and one case presented with lethargy and anorexia. The median age of the dog and cats in the lymphoma group was 9 years (range, 1–12 years). The dogs’ breed was Shiba while cats' breeds were Mixed and Ragamuffin.

In the carcinoma cases, peritoneal lesions were identified in the visceral peritoneum in one case (1 of 4; 25%) and in both parietal and visceral regions in three cases (3 of 4; 75%). All lesions were evenly distributed across the peritoneal regions. In the mesothelioma cases, peritoneal lesions were identified in the parietal region in all cases (5 of 5; 100%). In four cases (4 of 5; 80%), the parietal lesions were located on the diaphragm. In lymphoma cases, peritoneal lesions were identified in the parietal region for two cases (2 of 3; 67%) and the visceral region in one case (1 of 3; 33%). All lesions were evenly distributed across the peritoneal regions. The location was a characteristic showing significant differences among the lesions (P = 0.006). In terms of tumour shape, a localised mass was identified in three carcinoma cases (3 of 4; 75%), three mesothelioma cases (3 of 5; 60%) and none of the lymphoma cases (0 of 3; 0%). With regard to tumour shape, no significant differences were observed among the tumours (P = 0.22). For lesion enhancement patterns during the arterial, portal and equilibrium phases, heterogeneous enhancement was observed in three carcinoma cases (3 of 4; 75%), three mesothelioma cases (3 of 5; 60%) and one lymphoma case (1 of 3; 33%). For tumour enhancement patterns, no significant differences were observed across postcontrast phase images (P = 0.77). Lymphadenopathy was detected in only one carcinoma case (1 of 4; 25%), involving the sternal lymph node. Calcification was detected in only one carcinoma case (1 of 4; 25%). For lymphadenopathy and calcification, no significant differences were observed across tumours (P = 0.58). Ascites was detected in two carcinoma cases (2 of 4; 50%), five mesothelioma cases (5 of 5; 100%) and two lymphoma cases (2/3; 67%). Of the ascites, no significant differences across tumours (P = 0.27) were observed. The qualitative CT features are summarised in Table 1.

Table 1.

Qualitative computed tomography (CT) features among each peritoneal tumours including carcinoma, mesothelioma and lymphoma in five dogs and seven cats

CT features			Number and frequency of peritoneal tumour			p
			Carcinoma	Mesothelioma	Lymphoma
			N = 4	N = 5	N = 3
Location	Parietal		0/4 (0%)	5/5 (100%) a	2/3 (67%)	0.006
	Visceral		1/4 (25%)	0/5 (0%)	1/3 (33%)
	Parietal and visceral		3/4 (75%)	0/5 (0%)	0/3 (0%)
Tumour shape	Mass		3/4 (75%)	3/5 (60%)	0/3 (0%)	0.22
	Thickening		1/4 (25%)	2/5 (40%)	3/3 (100%)
Patterns of contrast enhancement	Arterial	Homo	1/4 (25%)	2/5 (40%)	2/3 (67%)	0.77
		Hetero	3/4 (75%)	3/5 (60%)	1/3 (33%)
	Portal	Homo	1/4 (25%)	2/5 (40%)	2/3 (67%)	0.77
		Hetero	3/4 (75%)	3/5 (60%)	1/3 (33%)
	Equilibrium	Homo	1/4 (25%)	2/5 (40%)	2/3 (67%)	0.77
		Hetero	3/4 (75%)	3/5 (60%)	1/3 (33%)
Lymphadenopathy	Presence		1/4 (25%)	0/5 (0%)	0/3 (0%)	0.58
	Absence		3/4 (75%)	5/5 (100%)	3/3 (100%)
Ascites	Presence		2/4 (50%)	5/5 (100%)	2/3 (67%)	0.27
	Absence		2/4 (50%)	0/5 (0%)	1/3 (33%)
Calcification	Presence		1/4 (25%)	0/5 (0%)	0/3 (0%)	0.58
	Absence		3/4 (75%)	5/5 (100%)	3/3 (100%)

^a Four cases (80%) were detected at diaphragmatic side of the parietal peritoneum.

In carcinoma cases, median tumour attenuation values in HU were 35.9 (range: 33.5 to 44.7) for the precontrast phase, 54.5 (range: 50.5 to 70) for the arterial phase, 102 (range: 81.3 to 103) for the portal phase and 122.3 (range: 83.3 to 129.5) for the equilibrium phase. For carcinoma cases, an overall significant difference (P = 0.007) was observed; however, post hoc tests failed to confirm significant differences across phase images.

For mesothelioma cases, the median tumour attenuation values HU were 34.7 (range: 29.6 to 44.6) for the precontrast phase, 60.7 (range: 31.7 to 128.3) for the arterial phase, 144.3 (range: 89 to 172.3) for the portal phase and 104.3 (range: 84.7 to 139.3) for the equilibrium phase postcontrast images. These mesothelioma attenuation values demonstrated significant differences (P = 0.003). Post hoc tests indicated a significant and high mesothelioma attenuation value during the portal (P = 0.04) and equilibrium phases (P = 0.04) compared to the precontrast phase.

For lymphoma cases, median tumour attenuation values were 31.7 (range: 31.3 to 48.7) HU for the precontrast phase, 66.3 (range: 57.3 to 76.7) for the arterial phase, 78 (range: 75.3 to 89.3) for the portal phase and 91.7 (range: 84.7 to 110.3) for the equilibrium phase. Within lymphoma cases, significant differences (P = 0.04) were observed, but post hoc tests failed to verify this across phase images. Tumour attenuation values for precontrast and postcontrast phases are shown in Figure 1. Representative images of each peritoneal tumour are shown in Figure 2.

Figure 1.

Attenuation values of peritoneal tumours, including carcinoma (A), mesothelioma (B) and lymphoma (C) in precontrast, arterial phase, portal phase and equilibrium phase postcontrast images. The attenuation value of mesothelioma was significantly high in the portal phase (P = 0.04) and equilibrium phase post‐contrast (P = 0.04) compared to pre‐contrast images. * P < 0.05.

Figure 2.

Representative axial portal CT images of carcinoma (A), mesothelioma (B) and lymphoma (C). Carcinoma identified in visceral or both parietal and visceral peritoneum (A, arrow). Mesothelioma identified in the diaphragmatic side of the parietal peritoneum with delayed enhancement (B, arrow). Peritoneal lymphoma presented as a thickening lesion (arrowhead) without lymphadenopathy.

Discussion

In humans, peritoneal carcinoma is predominantly the result of peritoneal dissemination from ovarian cancer. ¹⁹ , ²⁰ Therefore, a diagnosis of primary peritoneal carcinoma is made by ruling out ovarian cancer. ¹⁹ In veterinary medicine, primary peritoneal carcinoma is rare, and in cats, secondary peritoneal carcinomatosis is commonly associated with pancreatic adenocarcinoma, small intestinal adenocarcinoma and hepatocellular carcinoma. ³ In this study, four cats with carcinoma were suspected to have primary peritoneal carcinoma, as no primary tumour was identified on CT and there was no history of cancer. In cats, to the best of our knowledge, there is only one reported case of primary peritoneal carcinoma. ¹⁵ Although information is limited, CT examination of the cat with primary peritoneal carcinoma identified peritoneal thickening, fat stranding and calcification. ¹⁵ In humans, 30% of primary peritoneal carcinomas present with calcification and 82% with ascites. ²¹ Common findings of primary peritoneal carcinomas include peritoneal and omental nodules or caking, which typically show enhancement. ²¹ Less common findings include solid pelvic masses, lymphadenopathy and pleural effusions. ²¹ CT findings of primary peritoneal carcinoma often overlap with those of other secondary peritoneal carcinomatosis and lymphomatosis. ²¹ , ²² Therefore, further study of CT findings between primary and secondary peritoneal carcinoma in cats is needed. Although this study was limited by the small number of cases to establish clear trends, several carcinoma characteristics were observed. In this study, carcinoma was located in the visceral peritoneum or in both the parietal and visceral. On the other hand, mesothelioma and lymphoma were located in either the parietal or visceral peritoneum. Differential diagnoses in peritoneal carcinomatosis include pseudomyxoma peritonei, mesothelioma, peritoneal lymphomatosis, tuberculosis and diffuse peritoneal leiomyomatosis. ²³ Thus, peritoneal tumour location may assist in the diagnosis of peritoneal carcinoma.

Mesothelioma often presents without imaging abnormalities. When abnormalities are present, the most common findings are serosal thickening, nodules and masses. ⁶ In this study, all cases with mesothelioma appeared as localised masses in the parietal peritoneum. And 80% of mesothelioma cases were identified on the diaphragm. The presence of a parietal pleural lesion without a visceral pleural lesion is a useful finding to differentiate pleural mesothelioma from other pleural malignant tumours. ⁵ Where lesions are identified, lesions on the diaphragm may facilitate differentiating peritoneal mesothelioma. In humans, mesothelioma macroscopic features are similar to those of peritoneal carcinomatosis, including ascites, diffuse and/or nodular thickening of the peritoneal serous membrane, invasion of the greater omentum, sometimes with the formation of omental cakes, and mesenteric masses. ²³ In dogs, mesotheliomas are classified into epithelial, fibrous (sarcomatoid) and biphasic (mixed) histological types. ²⁴ The epithelial type resembles carcinomas and adenocarcinomas, ²⁴ with histological type potentially influencing CT findings. In our study, the attenuation value of mesothelioma was significantly high in portal and equilibrium phase images compared to precontrast images, indicating delayed enhancement. Unfortunately, mesothelioma histological types were not assessed in this study. Further study is needed to assess relationships between the attenuation values of mesothelioma among each phase and histological type.

In humans, peritoneal lymphoma is thought to spread through the peritoneal folds and ligaments, originating from alimentary lymphoma or via lymphatic dissemination. ²⁵ In cats, B‐cell alimentary lymphoma is associated with peritoneal lymphomatosis. ²⁶ On US, common findings of feline peritoneal lymphomatosis include an alimentary mural mass lesion with loss of wall layering, accompanied by thickened, hypoechoic plaques or sheets of abnormal tissue spreading throughout the mesentery or omentum, along with a small volume of peritoneal effusion. ²⁶ In humans, peritoneal lymphomatosis resembles carcinomatosis in many aspects, the most helpful distinguishing feature of lymphomatosis being the presence of lymphadenopathy, a large gastrointestinal mass without wall layering and splenomegaly. ²⁷ In this study, no alimentary masses or lymphadenopathy were observed. On CT, alimentary lymphoma typically presents as a mass form and is associated with lymphadenopathy. ¹¹ Unfortunately, the lymphoma cases did not show CT findings or clinical symptoms indicative of gastrointestinal lymphoma, thus no gastrointestinal tract biopsies were performed. As the peritoneum is devoid of lymphoid tissue, the mechanism of invasion is unknown. ²⁸ Although this study does not identify an obvious primary lesion, lymphomas may arise outside peritoneal lesions. Further study is required to assess relationships between peritoneal lymphoma and the primary lesion. Our study suggests including peritoneal lymphoma in the differential diagnosis, even in the absence of mass and lymphadenopathy.

Our results suggest that CT may identify the peritoneal tumours. Although the sites of lesion occurrence overlap, mesothelioma may be differentiated based on tumour location and attenuation value. When a definitive diagnosis cannot be achieved by luminal fluid cytology, CT examination may be a useful diagnostic approach. In addition, our findings may provide useful indicators for identifying lesions during ultrasonographic examination.

This study has several limitations. First, only a small number of peritoneal tumours were included. The small number of cases may influence statistical results in the attenuation value of adenocarcinomas and lymphomas. Therefore, the results of this study should be interpreted as preliminary findings. Second, this study was retrospectively designed. Multiple peritoneal lesions were present in each tumour, and individual histopathological diagnoses were not performed for all lesions. Therefore, this study contains potential inaccuracies in tumour classification. Because the tissue was not evaluated in all cases, the CT findings may not accurately reflect the lesion. Third, a fixed time‐delay method was used. Contrast‐enhanced techniques include fixed time‐delay or bolus tracking method. ²⁹ The fixed time‐delay method results in less consistent contrast enhancement compared with bolus tracking. ²⁹ Therefore, further studies are needed to evaluate the relationship between attenuation value and contrast enhancement techniques. Fourth, this study consisted of dogs and cats. CT findings may differ between dogs and cats. Because it may not reflect species‐specific CT findings, further studies are needed to evaluate CT findings in a larger population of dogs and cats with peritoneal tumours.

In conclusion, carcinoma was located in either visceral or both parietal and visceral peritoneum. Mesothelioma tended to localise on the diaphragm and showed delayed enhancement. Finally, peritoneal lymphoma was not associated with lymphadenopathy.

Statement of consent

The pet owners described in this study provided informed consent for the diagnostic procedures, treatment and use of clinical data such as medical history, imaging studies and histopathological findings for research and publication purposes. As each diagnostic study and initiated treatment were part of daily clinical activities, the study did not reach the threshold for submission to the local ethics and welfare committee.

Conflicts of interest and sources of funding

The authors declare no conflicts of interest or sources of funding for the work presented here.

Data availability statement

Data supporting the findings of this study are available from the first author, [T.T.], upon reasonable request.