Keywords incidentaloma - methods & techniques - ultrasound - CEUS - diagnosis - splenic lesion
Introduction
In ancient times, the spleen was called the “organum plenum mysterii” before its function
and significance were known [1 ]. Today, the spleen still does not receive much attention when evaluating imaging
of the abdomen and is called the “forgotten organ” in the abdomen [2 ], since focal splenic pathologies are rare with a world population incidence of only
up to 0.2% [3 ]. In this context, incidental focal lesions of the spleen present a particular problem
in everyday clinical practice, because an investigator’s personal experience with
splenic pathologies is likely to be limited.
Due to a general reluctance to perform splenic biopsy, costly multimodality imaging
is often requested for tumor characterization [4 ]. In the liver, contrast-enhanced ultrasound (CEUS) is the primary standard procedure
for hepatic incidentalomas [5 ]. However, data are limited regarding the value of CEUS for the evaluation of incidental
splenic lesion malignancy [6 ]. The aim of the present study was to evaluate B-mode ultrasound (B-US) and arterial
CEUS perfusion patterns of focal splenic incidentalomas (FSIs), and to correlate US
patterns with benignity and malignancy via histologic examination and/or clinical
course.
Patients and methods
Within the recruitment phase between 2004 and 2021, all patients with an FSI detected
by B-US were investigated additionally with CEUS at our Interdisciplinary Center of
Ultrasound Diagnostics (a tertiary healthcare facility at a university hospital),
which is an ultrasound reference center for splenic pathologies. All lesions were
examined by a single German Society for Ultrasound in Medicine (DEGUM) Level-III qualified
examiner with more than 35 years of experience in the field of abdominal sonography
(C.G. internal medicine) [7 ]
[8 ]. FSIs were defined as asymptomatic and unexpected splenic lesions discovered incidentally
on B-US, unrelated to the presenting illness, according to the World Federation for
Ultrasound in Medicine and Biology position paper on incidental splenic findings [6 ]. Wedge-shaped subcapsular lesions with absent enhancement on CEUS that were diagnosed
as splenic infarcts were not included in the study [9 ]. Furthermore, purely cystic anechoic lesions were not included in this study, because
CEUS is not indicated in these cases.
During the specified period, approximately 285,000 sonographic examinations were performed
in our ultrasound center, and 174 FSIs were found. The prevalence of splenic incidentalomas
was approximately 0.06%. This study was approved by the local ethics committee and
conducted in accordance with the amended Declaration of Helsinki. Informed consent
for the US examination was obtained from each patient.
The inclusion criteria were 1) a solid round splenic lesion, an echogenic cystic splenic
lesion, or a cystic splenic lesion with solid parts; 2) standardized documentation
of B-US and CEUS examinations; 3) no relationship between the splenic lesion and the
presenting illness of the patient, for which the investigation was performed; and
4) confirmation of the diagnosis of FSI by histologic examination and/or clinical
and radiological follow-up.
In total, 35/174 patients (20.1%) with an FSI were excluded: 32/35 (91.4%) due to
the absence of diagnostic confirmation regarding malignancy and benignity and 3/35
(8.6%) due to the absence of standardized documentation of US examinations. Finally,
data from 139 patients with an FSI were analyzed retrospectively.
Classification of splenic incidentalomas based on medical history
According to the clinical background of the patients, the lesions were divided into
two groups [10 ]:
Splenic incidentalomas in a strict sense : splenic lesions as incidental findings without known malignant disease;
Splenic incidentalomas in an extended sense : splenic lesions as incidental findings in patients with a prior history or current
evidence of a malignant disease.
Ultrasound examination
The B-US examinations were performed with an Acuson Sequoia 512 GI ultrasound machine
(Siemens, Germany) and a 4C1 curved-array transducer with a frequency of 4 MHz.
The CEUS investigations were conducted with the same transducer in contrast-specific
mode (1.5 MHz) and in accordance with the European Federation of Societies for Ultrasound
in Medicine and Biology (EFSUMB) guidelines [7 ]. A bolus injection of 2.4 ml of the contrast medium SonoVue (Bracco Imaging S.p.A.,
Milan, Italy) was administered via peripheral venous access. This was followed by
10 ml of NaCl 0.9%. For the first 30 seconds, the perfusion patterns of the lesions
were continuously examined and recorded in a video clip. Subsequently, several short
examinations were performed at 1-minute intervals for up to 3 minutes, and the changes
in the perfusion pattern were saved as images. In patients with multiple lesions,
the largest lesion was selected as the reference lesion. The following B-US and CEUS
data were evaluated retrospectively.
B-mode ultrasound
The echogenicity of the lesions was classified as hypoechoic, hyperechoic, echogenic
cystic, or complex (cystic with solid parts) compared with the echogenicity of splenic
parenchyma used as an in vivo reference [6 ]
[11 ]
[12 ].
The number of lesions was classified as solitary or multiple.
The size of lesions was measured in centimeters (maximum diameter).
Contrast-enhanced ultrasound
1. The arterial enhancement (AE) of FSIs was categorized as hyperenhancement, isoenhancement,
hypoenhancement, or absent enhancement [2 ]
[6 ]
[11 ]
[13 ]. The arterial phase was defined as the period from the earliest arrival of the contrast
agent at the spleen until 60 seconds thereafter. Splenic tissue was considered as
an in vivo reference to evaluate the AE of the contrast agent [14 ].
The B-US and CEUS data were evaluated retrospectively by two independent, experienced
investigators (C.G., E.S.). In the event of discrepancies, the final decision was
made by a third experienced investigator (H.F.).
Classification of splenic lesions based on B-mode ultrasound and contrast-enhanced
ultrasound patterns
After the B-US and CEUS examinations, the lesions were classified into five groups
according to echogenicity on B-US (hypoechoic, hyperechoic, echogenic cystic, or complex)
and arterial enhancement on CEUS (hyperenhancement, isoenhancement, hypoenhancement,
or absent enhancement). The classification criteria were defined according to the
modified classification of Bert and Görg et al. [11 ] ([Table 1 ]).
Table 1 Classification of incidental focal splenic lesions in the study patients (N = 139).
B-US: B-mode ultrasound; CEUS: contrast-enhanced ultrasound
Group I
A
Hypoechoic on B-US and arterial hyperenhancement on CEUS
B
Hypoechoic on B-US and arterial iso-/hypoenhancement on CEUS
Group II
A
Hyperechoic on B-US and arterial hyperenhancement on CEUS
B
Hyperechoic on B-US and arterial iso-/hypoenhancement on CEUS
Group III
Echogenic cystic or complex on B-US and absent enhancement on CEUS
Statistical analysis
Statistical evaluation was performed on the categorical variable using Fisher’s exact
test and on continuous data using Mann–Whitney tests. Cohen’s kappa statistics were
applied to measure interrater reliability, and a p -value of <0.05 was defined as significant.
Results
Demographic and clinical data
Of the 139 patients, 73 were men and 66 were women. The average age was 56.5 years,
with a range of 14–84 years. The final diagnosis was malignant FSI (mFSI) in 9/139
cases (6.5%) and benign FSI (bFSI) in 130/139 cases (93.5%). In total, the diagnosis
was made by histologic confirmation in 18/139 cases (12.9%; 10 biopsies, 5 splenectomies,
3 autopsies) and by clinical and/or radiological follow-up in the remaining 121/139
cases (87.1%). The average time of the follow-up was 5 years and 6 months.
In 7/9 malignant cases (77.8%) the diagnosis of an mFSI was confirmed by histologic
examination of the FSI. In the other 2 cases (22.2%), the diagnosis was confirmed
by a histologic examination of a distant malignant lesion and by complete regression
of FSI under chemotherapy. Of 130 bFSIs, the diagnosis was confirmed based on histologic
examination in 11/130 cases (8.5%) and on clinical and radiological follow-up in 122/130
cases (91.5%). The final diagnoses of FSI in all the patients are shown in [Table 2 ].
Table 2 Diagnosis of focal splenic incidentalomas in all study patients.
Benign FSI
Number (%) of patients
Malignant FSI
Number (%) of patients
AML: acute myeloid leukemia; FSI: focal splenic incidentaloma. *The diagnosis was
confirmed histologically.
Indeterminate benign masses
119 (91.5)
Malignant splenic lymphoma *
7 (77.8)
Splenoma (splenic hamartoma) *
3 (2.3)
Melanoma metastasis *
1 (11.1)
Granulomatous inflammation *
2 (1.6)
Chloroma in AML *
1 (11.1)
Non-specific inflammatory reaction *
1 (0.8)
–
–
Hemangioma *
1 (0.8)
–
–
Normal spleen tissue *
1 (0.8)
–
–
Splenic cyst *
1 (0.8)
–
–
Extramedullary hematopoiesis *
1 (0.8)
Littoral cell angioma *
1 (0.8)
Total number of patients with benign FSI
130 (100)
Total number of patients with malignant FSI
9 (100)
Classification of splenic incidentalomas based on medical history
In total, 97/139 patients (69.8%) had splenic incidentalomas in a strict sense. In
this group, 91/97 lesions (93.8%) were benign and 6/97 lesions (6.2%) were malignant.
The malignant lesions were nodular splenic lymphomas in all patients.
In addition, 42/139 patients (30.2%) had splenic incidentalomas in an extended sense,
with known malignant disease in their medical history (24 cases with a current and
18 cases with a prior history of malignant underlying disease). In this group, 39/42
lesions (92.2%) were benign and 3/42 lesions (7.1%) were malignant. In patients with
a current malignant underlying disease, 2/24 lesions (8.3%) were malignant, and, in
patients with a prior history of malignant underlying disease, 1/18 lesions (5.6%)
were malignant.
The malignant lesions were one melanoma metastasis in a patient with prior history
of malignant melanoma, one lymphoma in a patient with known lymphoma, and one chloroma
in a patient with known acute myeloid leukemia (AML).
Of these 42 patients with FSI in an extended sense, 30/42 cases (71.4%) had a non-hematologic
underlying malignancy (23 carcinomas, 2 germ cell tumors, 4 malignant melanomas, 1
malignant melanoma and carcinoma), 9/42 cases (21.4%) had a hematologic underlying
malignancy (6 malignant lymphomas, 3 cases of acute myeloid leukemia), and 3/42 cases
(7.1%) had both a non-hematologic and a hematologic underlying malignancy (3 malignant
lymphomas and carcinomas).
The frequency of mFSI was not significantly different in patients with a known malignant
disease compared with those without a known malignant disease (p > 0.99, Fisher’s exact test; [Fig. 1 ]).
Fig. 1 Malignancy rate of N = 139 splenic incidentalomas according to the presence of an
underlying malignant disease in the medical history of study patients.
Ultrasound examination
B-mode ultrasound data
On B-US, 71/139 FSIs (51.1%) were hypoechoic (group I) ([Fig. 2 ]), 58/139 (41.7%) were hyperechoic (group II) ([Fig. 3 ]), and 10/139 (7.2%) were echogenic or cystic/complex (group III), 7 of which were
echogenic cystic and 3 had a complex echogenicity ([Fig. 4 ]).
Fig. 2 Group 1: (A ) A 71-year-old male patient with an incidentally detected hypoechoic splenic lesion
on B-mode ultrasound and without a history of malignancy. (B ) On contrast-enhanced ultrasound, the lesion showed arterial hyperenhancement after
60 s. A clinical follow-up of 5 years and 3 months revealed no evidence of malignancy.
The diagnosis of an indeterminate benign mass was made. (C ) A 63-year-old male patient with an incidentally detected hypoechoic splenic lesion
on B-mode ultrasound (arrow) and without a history of malignancy. (D ) On contrast-enhanced ultrasound, the lesion showed an arterial isoenhancement after
27 s (arrow). The histopathologic examination of an enlarged mesenteric lesion revealed
the diagnosis of Hodgkin’s lymphoma. In the follow-up ultrasound after rituximab therapy,
the lesion was no longer detectable. The diagnosis of splenic involvement in Hodgkin’s
disease was made. (E ) A 63-year-old male patient with an incidentally detected hypoechoic splenic lesion
on B-mode ultrasound, without a history of malignancy. (F ) On contrast-enhanced ultrasound, the lesion showed arterial hypoenhancement after
37 s. A splenectomy was performed, and the diagnosis of primary splenic follicular
lymphoma was confirmed histologically.
Fig. 3 Group 2: (A ) A 54-year-old female patient with an incidentally detected hyperechoic splenic lesion
on B-mode ultrasound (arrow), without a history of malignancy. (B ) On contrast-enhanced ultrasound, the lesion showed arterial hyperenhancement after
9 s. A sonographic follow-up of 4 years and 6 months revealed no size progression.
The diagnosis of an indeterminate benign mass was made. (C ) A 56-year-old male patient with an incidentally detected hyperechoic splenic lesion
on B-mode ultrasound and with known acute myeloid leukemia (AML). (D ) On contrast-enhanced ultrasound, the lesion showed arterial hypoenhancement after
44 s. The patient died 1 year and 3 months later due to a cerebral manifestation of
AML and massive cerebral hemorrhage. At autopsy, focal nodular blast infiltration
(chloroma) of the spleen was histologically confirmed.
Fig. 4 Group 3: (A ) A 30-year-old male patient with an incidentally detected echogenic cystic splenic
lesion on B-mode ultrasound, without a history of malignancy. (B ) On contrast-enhanced ultrasound, the lesion showed absent enhancement during the
whole investigation. A sonographic follow-up of 8 years and 4 months revealed no size
progression. The diagnosis of an indeterminate benign mass was made. (C ) A 45-year-old female patient with an incidentally detected complex lesion (cystic
with solid parts) on B-mode ultrasound and prior known endometrial carcinoma. (D ) On contrast-enhanced ultrasound, the lesion showed predominantly absent enhancement
during the whole investigation. A sonographic follow-up of 3 years revealed no size
progression. The diagnosis of an indeterminate benign mass was made.
Of the malignant lesions, 6/9 cases (66.6%) were hypoechoic ([Fig. 2 ]C–D and E–F) and 3/9 (33.3%) cases hyperechoic on B-US ([Fig. 3 ]C–D).
Further detailed diagnostic data from the B-US examination are summarized in [Table 3 ].
Table 3 B-mode ultrasound data of the N = 139 patients, subdivided between final malignant and benign focal splenic incidentalomas.
B-US feature
All FSIs
N = 139
Malignant FSIs
n = 9
Benign FSIs
n = 130
p -value
B-US: B-mode ultrasound; FSI: focal splenic incidentaloma unless otherwise noted,
the values are indicated as number (%). A p -value of <0.05 was defined as significant. *Fisher’s exact test; **Mann-Whitney test
Hypoechoic
71 (51.1%)
6 (66.7%)
65 (50.0%)
0.49*
Not hypoechoic
68 (48.9%)
3 (33.3%)
65 (50.0%)
Solitary lesion
84 (60.4%)
7 (77.8%)
77 (59.2%)
0.48*
Multiple lesions
55 (39.6%)
2 (22.2%)
53 (40.8%)
Average size of lesions (cm)
2.6
3.8
2.6
0.17**
Contrast-enhanced ultrasound data
Regarding AE, 44/139 cases (31.7%) showed arterial hyperenhancement ([Fig. 2 ]B and [Fig. 3 ]B), 18/139 cases (12.9%) arterial isoenhancement ([Fig. 2 ]D), 67/139 cases (48.2%) arterial hypoenhancement ([Fig. 2 ]F), and 10/139 cases (7.2%) absent arterial enhancement ([Fig. 4 ]B and D). Of 9 mFSIs, 8/9 cases (88.9%) showed hypoenhancement and 1/9 cases (11.1%)
isoenhancement.
Regarding echogenicity in 5/139 cases (3.6%) and arterial enhancement in 7/139 cases
(5.0%), there was a discrepancy between the first and second investigator, and the
final decision was made by a third investigator. The agreement between the examiners
for the ultrasound finding was “very good” (Cohen’s kappa = 0.9).
Classification of splenic lesions based on B-mode ultrasound and contrast-enhanced
ultrasound patterns
The number and malignancy rate of FSIs according to the modified classification of
Bert and Görg et al. [11 ] is presented in [Fig. 5 ].
Fig. 5 The number and malignancy rate of 139 FSIs in different groups according to the modified
classification of Bert and Görg.
Discussion
Improved imaging techniques with high resolution and the increased use of additional
imaging methods have led to an increase in incidental findings in different organs
[15 ]. In the literature, the FSIs detected via imaging methods are described as rare,
with a prevalence of less than 1% [16 ]
[17 ]. In accordance with previous studies, the FSIs in this study were indeed rare and
revealed a prevalence of 0.06%. Compared with previous studies, the smaller number
of patients enrolled in the study during the long recruitment period may be due to
differences in the definitions of the inclusion criteria [10 ]
[18 ].
Overall, 6.5% of the FSIs were shown to be malignant, and 93.5% were benign. 6.2%
of lesions were malignant in patients without a history of malignancy, and 7.1% of
lesions in patients with a history of malignancy. Previous studies have shown a malignancy
rate of 1–33.8% of splenic lesions in patients without a history of malignancy and
33.8–86.7% in patients with a history of malignancy [10 ]
[18 ]
[19 ]. This wide range of malignancy rates in various studies may be caused by differences
in patient spectra and in the definition criteria of splenic incidentalomas ([Table 4 ]).
Table 4 Malignancy rates of splenic lesions in patients with and without a history of malignancy.
Imaging modality
Cases
Year
Author
Patients with a history of malignancy (%)
Patients without a history of malignancy (%)
Study characteristics
CT: computed tomography; MRI: magnetic resonance imaging; PET/CT: positron emission
tomography/computed tomography; US: ultrasound; *: these lesions were not defined
as incidentalomas
US
136
2011
Stang et al. [10 ]
80.9
33.8
The criteria for splenic incidentalomas were not defined
Cysts, infarction, abscess, sarcoidosis, and diffuse micronodular infiltration by
lymphoma were excluded from the study
CT, MRI, and PET/CT
53
2013
Dhyani et al. [19 ]
86.7
2.6
Only patients under the age of 30 years were included
Lesions detected on staging examinations performed for evaluation of malignant disease
were included in the study
Splenic infarcts were included
CT
379
2018
Siewert et al. [18 ]
33.8*
1.0
Splenic infarcts and calcified granulomas were excluded
Splenic cysts were included
Splenic lesions in patients with known malignancy were not considered to be incidentalomas
Only lesions in asymptomatic patients without history of malignancy were defined as
incidentalomas
US
139
2022
Present study
7.1
6.2
Focal splenic incidentalomas were defined as asymptomatic and unexpected splenic lesions
discovered incidentally on US, unrelated to the presenting illness
Splenic infarcts and purely cystic anechoic lesions were excluded
We observed no significantly different frequencies in malignancy of FSIs according
to the presence or absence of a malignant disease in the medical history of the patients
(p > 0.99, Fisher’s exact test). Based on these findings, known malignant disease should
not be considered as a highly determinant factor in the diagnostic workup of FSIs.
This result is in line with previous studies, which showed that the spleen is an uncommon
site for metastatic disease. A large autopsy study in 1898 patients with a solid malignant
tumor showed that splenic metastasis was present in only 3% of patients [20 ]. Furthermore, an ultrasound study in 680 patients with histologic evidence of malignant
lymphoma detected splenic involvement in less than 15% of patients [21 ].
Regarding B-US patterns, EFSUMB guidelines describe small echogenic lesions as usually
benign and hypoechoic lesions as more frequently malignant [7 ]. In this study, in accordance with EFSUMB statements, the majority of malignant
lesions (66.7%) were hypoechoic and larger than benign lesions (3.6 cm vs 2.6 cm).
However, in terms of B-US echo patterns, we found no significant difference between
benign and malignant FSIs (p > 0.05), which suggests non-specificity of B-US characteristics with respect to the
malignancy of FSIs.
Regarding CEUS patterns, EFSUMB guidelines describe absent enhancement or arterial
phase hyper-/isoenhancement as characteristic of benign lesions [7 ]. In accordance with EFSUMB statements, in the present study, all FSIs with absent
enhancement or arterial hyperenhancement were benign. However, contrary to the EFSUMB
guidelines, one lesion with arterial isoenhancement was shown to be a malignant lymphoma.
These results are in accordance with findings from a previous study that demonstrated
that 46.3% of splenic lymphomas show arterial isoenhancement [22 ]. Therefore, arterial isoenhancement should not be used as a predictor factor of
benignity.
In the additional evaluation, we classified splenic incidentalomas using established
B-US and CEUS criteria into five groups and subsequently followed them in terms of
malignancy during their clinical course. The hypoechoic and hyperechoic lesions with
arterial hyperenhancement (groups 1a and 2a) and the echogenic cystic or complex lesions
with predominantly absent enhancement (group 3) were all benign. Furthermore, arterial
iso-/hypoenhancement on CEUS may indicate FSI malignancy. However, the malignancy
rate is different between hypoechoic and hyperechoic FSIs. In the present study, 16.2%
of hypoechoic lesions with arterial hypo-/isoenhancement (group 1b) and 6.3% of hyperechoic
lesions with arterial hypo-/isoenhancement (group 2b) were malignant. Based on these
results, the lesions reveal different malignancy rates depending on the group, and
this classification may be helpful in further evaluation of FSIs.
There are some limitations to this study. These include the general limitations of
ultrasound examinations, which are characterized by high interobserver and interequipment
variability. Furthermore, the study was performed only in patients who were referred
to the Interdisciplinary Center of Ultrasound Diagnostics for the investigation of
abdominal pathologies. Therefore, selection bias cannot be excluded. Histologic confirmation
was not performed in all patients with an FSI. However, all diagnoses in these patients
were verified by clinical and/or radiological follow-up. Another limitation of our
study is the semiquantitative classification of the ultrasound data, which may lead
to more scope for interpretation than a quantitative method. However, the interrater
observer variability for the ultrasound findings demonstrated “very good” agreement.
Due to the retrospective nature of the study and the relatively small number of subjects
(N = 139), further prospective multicentric studies are needed to validate our findings.
Conclusion
In summary, classification according to the B-US and CEUS patterns may be useful in
evaluating the malignancy of FSIs. In all patients with a hypoechoic or hyperechoic
lesion with arterial hypo-/isoenhancement, further imaging and short-term imaging
follow-up or histologic confirmation are indicated even if the patient has no history
of a malignant disease. In patients with a hypoechoic or hyperechoic lesion with arterial
hyperenhancement and in those with an echogenic cystic or complex lesion with predominantly
absent enhancement, only imaging follow-up should be performed.
