In this case series, we systematically describe ultrasound features prevalent in African patients with KS. The most common findings were enlarged lymph nodes and effusions. In liver and spleen, focal lesions were observed; these were mostly hyperechoic, resembling hemangiomas. Mixed hyper-hypoechoic lesions (target- and bull’s eye lesions) were also seen. In three patients, we documented a “sponge-like pattern” of spleen tissue. Findings in spleen and liver were often faint and better visible using the linear probe; the use of the high-frequency probe likely added to the relatively high proportion of solid organ involvement documented in our series. The sonographic findings identified resemble those previously described in pictorial reviews [5, 6]; to our knowledge, the frequency of the various findings has not been previously described.

Ultrasound findings of the FASH scan—effusions, abdominal lymphadenopathy and focal spleen lesions—are frequently used to aid the diagnosis of disseminated TB in resource-limited high-prevalence settings. It is, therefore, important to discuss the overlap with findings in KS patients to prevent misdiagnosis.

The enlarged lymph nodes resemble those of disseminated TB, although the inguinal localization is more suggestive of KS. Nevertheless, abdominal lymphadenopathy due to KS can also be found and can be mistaken for a sign of disseminated TB. The KS-associated effusions in our patients showed no differentiating patterns and could also be mistaken for a sign of disseminated TB. In the interpretation of the findings, it should be noted that KS-associated effusions and intra-abdominal lymphadenopathy, have a broad differential diagnosis, including extra-pulmonary TB, disseminated Mycobacterium avium complex (MAC) infection, HIV lymphadenopathy, Hodgkin- or non-Hodgkin lymphoma, and other human herpes virus 8 (HHV-8)-associated illnesses, such as multicentric Castleman disease (MCD) and primary effusion lymphoma [8, 9].

The sonographic lesions of KS in spleen and liver are more pathognomonic. Hyperechoic lesions in a patient with HIV, a low CD4 T-cell count and systemic symptoms, are highly suggestive of KS and are unlikely to be mistaken. They should prompt an immediate (re-)evaluation of the entire skin, especially of legs and groins, and oral mucosa to search for KS skin manifestations. We also identified hypoechoic splenic lesions with a hyperechoic center (target lesions), and these can be difficult to distinguish form micro-abscesses when they are small and the examiners less experienced.

Finally, we found a sponge-like pattern of the spleen, which has only recently been described in adult HIV positive patients [10]. This sponge pattern can be mistaken for spleen micro-abscesses, although the hypoechoic structures are generally smaller [1–2 mm] then lesions caused by TB (> 5 mm). Our data presented here suggest that the sponge pattern may be seen in about 10% of KS patients, although our absolute patient number was small (95% confidence interval: 0–21%). The pattern was originally described in association with a variety of opportunistic illnesses, including KS, disseminated MAC infection, pneumococcal infection, Hodgkin lymphoma and MCD as well as with HIV itself—still almost half of the reported cases were associated with HHV-8 associated diseases, such as KS or MCD [10]. A recent radiological review states that KS tends to “surround the arteries of the Malpighian corpuscles and has a stringy appearance”[4]—which could explain the sponge pattern in ultrasound—but actual histo-pathological evidence for this in newer literature is scarce. Still, an autopsy series from the early 1980s before HIV testing became available found KS involvement in 32 of 44 spleens of patients who died with clinical AIDS and reported that “in the spleen the perivascular white pulp is most commonly and extensively involved” [11].

In summary, our results highlight the importance to place ultrasound findings in clinical context and perform a detailed history (e.g., detailed symptoms, exposure to TB) and physical examination to assess for KS skin lesions. Only 1–2% of KS patients present with enlarged LN without skin lesions [12], thus in the vast majority of cases skin changes will be found if thoroughly looked for. All patients in our cohort showed typical skin changes. Our findings also underline that even in resource-limited settings a pathological confirmation is desirable for many diseases due to overlapping findings of various differentials; nevertheless this is rarely achievable and clinical observation of treatment response remains a mainstay of making diagnoses in our setting.

Our study was limited by the fact that it was performed in a single center with a relatively small sample size. In addition, while the diagnosis of KS was made clinically and supported histologically, we were not able to entirely rule-out diagnosis of other opportunistic infections or of concomitant TB (beyond clinical assessment, chest X-ray and urine LAM tests). Our clinic setting was also not conducive to sonographic follow-up of patients; therefore, we cannot report on the changes of the findings under treatment. However, to the best of our knowledge, this is the first systematic African case series providing estimates of frequency of sonographgic findings prevalent in Kaposi sarcoma. Overall ultrasound findings in KS seem to be more frequent than previously thought, especially if linear transducers are used to evaluate solid organs.