Keywords
Accuracy - anorectal - diagnostic - melanoma - positron emission tomography/computed
tomography
Introduction
Melanomas arise from melanocytes, which are derived from embryonic neural crest cells.
Anorectal melanoma is a rare, aggressive malignancy and most arises from the dentate
line.[1] They account for <1% of all malignant melanomas and fewer than 3% of all anal tumors.[1],[2],[3] The incidence of anorectal melanoma is about 1–2 cases per million population per
year in the United States, with statistical estimates in similar range in Europe and
Asia.[4] The incidence of anorectal melanoma is thought to be increasing.[4] Most of the patients present with nonspecific symptoms and as a result tend to have
disease progression to advanced stage at initial presentation.[1] The overall risk factors for anorectal melanoma are also poorly known. The relationship
between cutaneous melanoma and ultraviolet B light, which is a carcinogenic stimulus,
is well documented.[5] However, it is not obvious what triggers the development of anal melanoma as anal
mucosa is never exposed to sunlight. As with cutaneous melanomas, which are 20 times
more common in whites than blacks, there is no such evidence for anal melanomas.[3] Tumor stage is an independent prognostic variable for overall survival. However,
there are no well-established staging systems and treatment protocols due to its rarity.
Siegal et al. reported that survival for Stage I patients (local disease) was 48 months,
Stage II (local disease with regional nodes) was 12 months, and Stage III (distant
metastases) was 10 months.[6] Thus, it is important to have accurate imaging modality for staging anorectal melanoma.
The role of positron emission tomography/computed tomography (PET/CT) in staging and
restaging of several cancers has already been established. The literature regarding
the role of PET/CT in anorectal melanoma is sparse as it is a rare type of tumor.
Only few case reports and case series are documented in literature. No studies or
any guidelines establishing the role of PET/CT in anorectal melanoma are currently
published. The aim of our retrospective audit was to report our experience in evaluating
the diagnostic accuracy of 18F-fluorodeoxyglucose (18F-FDG) PET/CT in staging and restaging of anorectal melanoma.
Materials and Methods
This was an institutional review board-approved, single-institution, retrospective
observational study. Consecutive patients were included. Inclusion criteria were biopsy-proven
cases of anorectal melanoma referred for initial staging and for suspected recurrence/restaging
who underwent a PET/CT between January 2006 and December 2015. Patients in whom biopsy
report was not available or with synchronous malignancy at any other site were excluded
from the study. Diagnostic performance of PET/CT was evaluated for baseline staging
and restaging with clinical suspicion of recurrence. PET/CT studies were evaluated
for the ability to detect the primary disease, nodal, and other sites of metastases.
Histopathology was used as gold standard. When tissue diagnosis was not available,
clinical or radiological follow-up was used as reference standard. The sensitivity
(SN), specificity (SP), positive predictive value (PPV), negative predictive value
(NPV), and accuracy were calculated at 95% confidence interval. At present, as there
is no staging system specific to the disease, a clinical staging as suggested by several
retrospective studies for mucosal melanomas was used, Stage I as localized disease
only and Stage II as regional lymph node (LN) involvement. Nodes in the pelvis below
the common iliac were considered locoregional nodes. Disease in nodes from common
iliac and above and other organs was Stage III.
Patient preparation and positron emission tomography/computed tomography protocol
All patients were asked to fast for at least 6 h prior to the study and blood glucose
levels checked and confirmed to be below 150 mg/dl (8.3 mmol/L). The studies were
taken 60 ± 15 min following intravenous administration of 5 MBq/kg body weight of
18F-FDG. Imaging was performed using Philips Gemini TF TOF 16/64 PET/CT scanners (PET
crystal-LYSO). After obtaining a scout image, breath-hold CT was acquired followed
by whole-body CT and the PET acquisition in the same anatomic location. Whole-body
CT was acquired from skull base to midthigh region (slice thickness 2 mm, 120 kV voltage,
automated mA, rotation time 0.5 s, pitch 0.83, image matrix 512 × 512, field of view
(FOV) 600 mm) without any breath-hold instructions. Separate breath-hold CT was acquired
(slice thickness 3 mm, 120 kV, automated mA, pitch 1.08, FOV 300 mm, image matrix
512 × 512) for the evaluation of the lungs. Eighty milliliters of oral contrast was
administered in all eligible patients at a rate of 1.8 ml/s and scan delay of 50 s.
Water-based oral contrast was given for bowel distension. Contrast-enhanced CT was
used for diagnostic purpose as well as for attenuation correction of PET data. PET
was acquired in three-dimensional mode in about 5–8 bed positions with an acquisition
time of 60–90 s/bed position, axial FOV 576 mm, and in plane spatial resolution of
4 mm.
Image reconstruction and interpretation
The images were reconstructed iteratively using RAMLA algorithm. CT attenuation correction,
dead time correction, and decay correction were applied. Images were viewed on Philips
workstation having extended brilliance workspace version 4.5.3.40140, equipped with
fusion viewer software that enables the display of PET, CT, and fusion PET/CT images.
The images were reviewed by two experienced nuclear medicine physicians. Any area
with intensity greater than background that could not be identified as physiological
activity on PET images or which on CT correlation did not fit into benign (infective/inflammatory/degenerative)
was considered to be suggestive of tumor on PET/CT study. CT criteria for positive
metastatic node included size more than or equal to 1 cm in short-axis measurement,
rounded nodes, loss of fatty hilum, central necrosis, and/or contrast enhancement.
Morphologic characteristics of pulmonary lesions on CT were taken positive even in
the absence of 18F-FDG uptake.
The diagnostic accuracy of 18F-FDG PET/CT for the detection of primary tumor/recurrence and the metastatic lesions
was correlated with either histopathology or subsequent clinical/follow-up imaging.
Statistical analysis
All the PET/CT studies were evaluated for the ability to detect the primary disease,
nodal, and distant metastatic sites. The SN, SP, PPV, NPV, and accuracy were calculated
at 95% confidence interval.
Results
Initial staging
A total of 61 (45 males and 16 females) patients with an age range of 28–88 years
were included for initial staging. Among the 61 patients, 13.1% had Stage I, 29.5%
had Stage II, and 57.4% had Stage III at initial presentation. Primary tumor site
was localized to anal canal, rectum, and anorectal region [[Table 1]]. The primary tumor/lesion was detected in 57 patients out of 61. In four patients
in whom no lesions were detected, PET/CT was true negative in two and false negative
[Figure 1] in the other two patients (SN, 96.6%; SP, 100%; PPV, 100%; NPV, 50%; and accuracy,
96.7%). Regional nodal involvement was correctly detected in 46 out of 61 patients
with false negative in one patient (SN, 97.9%; SP, 100%; PPV, 100%; NPV, 93.3%; and
accuracy, 98.4%). Histopathological correlation was available for primary and regional
nodal disease in all patients [Figure 2]. Twenty-two patients had nonregional nodes positive on PET/CT, whereas distant metastases
were detected in 25 patients with SN 100%, SP 100%, PPV 100%, and NPV 100% [Figure 3]. These were confirmed with subsequent follow-up imaging [[Table 2]].
Figure 1: A 52-year-old patient presented with bleeding per rectum. Biopsy revealed
features consistent with malignant melanoma. FDG PET/CT was done as baseline imaging.
MIP image (a) shows no abnormal focal increased tracer uptake in the pelvic region.
No focus of hypermetabolism or soft tissue lesion is seen in the anal canal region.
Postsurgery histopathology results revealed submucosal malignant melanoma of rectum
with reactive regional nodes. Thus, PET/CT showed false-negative results at the primary
tumor site. MIP (b) and fused (c) images show focal uptake in the reactive right hilar
node. FDG: Fluorodeoxyglucose; PET/CT: Positron emission tomography/computed tomography;
MIP: Maximum Intensity Projection
Figure 2: A 41-year-old patient presented with blood in stools, pain in perianal area,
and weight loss. Biopsy revealed malignant tumor which had epithelioid and slightly
spindle-shaped appearance. Immunohistochemistry was positive for HMB 45 and S100 proteins.
MIP image (a) shows increased FDG tracer uptake in the pelvis and below it. Fused
transaxial PET/CT images reveal FDG-avid growth in the anal canal (b) with metastatic
right internal iliac node (c) and perirectal nodule (d). FDG: Fluorodeoxyglucose;
PET/CT: Positron emission tomography/computed tomography; MIP: Maximum Intensity Projection
Figure 3: A 33-year-old patient presented with bleeding per rectum. Rectal growth
biopsy is suggestive of malignant melanoma. Immunohistochemistry revealed tumor positive
for HMB 45 and S100 protein. FDG PET/CT scan was done for initial staging. MIP images
(a) show increased tracer uptake at multiple sites in the pelvis, liver, and multiple
vertebrae. Fused transaxial PET/CT images show large anorectal mass (b), metastatic
left internal iliac node (c), bilateral inguinal nodes (d), metastatic liver lesions
(e), skeletal lesions (f), and lung nodules(g). The patient was considered for palliative
chemoradiotherapy. FDG: Fluorodeoxyglucose; PET/CT: Positron emission tomography/computed
tomography; MIP: Maximum Intensity Projection
Table 1: Patient demographics
Table 2: Result of Initial staging
Restaging
A total of 24 (15 males and 9 females) patients were included for suspected recurrence/restaging
[[Table 3]]. All patients underwent surgery. 79.2% (19/24) of patients underwent surgery alone;
16.7% (4/24) of patients were treated with surgery followed by chemotherapy; and 4.2%
(1/24) of patients underwent surgery followed by radiotherapy. PET/CT detected disease
recurrence in 20 (83.3%) patients. Ten patients had recurrence at the primary site
[Figure 4], 8 of whom also had distant metastases [Figure 4a] and 2 had only locoregional metastatic nodes. In the remaining 10 patients, there
was no primary site recurrence; however, 2 patients had locoregional nodal and distant
metastases and 8 patients had only distant metastases. Histopathological confirmation
was available in 6 patients, whereas imaging and clinical follow-up confirmed recurrence
in the remaining patients. Thus, primary site recurrence was noted in 10 (41.7%) cases,
locoregional nodes in 10 (41.7%) cases, and distant metastases in 18 (75%) cases.
Liver was the most common site of distant metastases in 45.8% (11/24) of patients,
followed by lungs 37.5% (9/24). PET/CT was negative for recurrence in three patients.
This was confirmed by follow-up imaging studies in two cases and short-term clinical
follow-up in one case. PET/CT was false negative in one patient, which missed liver
metastases [Figure 5]. Ultrasonography-guided fine-needle aspiration cytology (FNAC) of the liver lesion
was done prior to the PET/CT scan, which had revealed metastatic malignant melanoma
consistent with known primary in anorectum.
Table 3: Patient demographics
Figure 4: A 89-year-old patient presented with biopsy proven anal malignant melanoma.
PET/CT scan was for restaging in the view of growth detected on per rectal examination.
MIP image (a) shows focal increased tracer uptake in the pelvis, liver, right lung,
and dorsal vertebra. Fused transaxial FDG PET/CT images show anorectal soft tissue
mass (short arrow) and left inguinal node (long arrow) (b); pararectal nodule (c);
liver metastases (d), metastatic dorsal vertebral lesion (e); and lung nodules (f).
The patient was considered for palliative chemotherapy. FDG: Fluorodeoxyglucose; PET/CT:
Positron emission tomography/computed tomography; MIP: Maximum Intensity Projection
Figure 5: A 53-year-old patient initially presented with pain in abdomen and underwent
anal polypectomy. Biopsy report revealed anorectal melanoma with immunohistochemistry
results positive for HMB 45 and S100 protein. Follow-up USG done revealed left liver
lobe hypoechoic suspicious lesion. USG done six months later showed increase in the
size of suspicious left liver lobe lesion. FNAC revealed metastatic malignant melanoma
consistent with known primary in the anorectum. FDG PET/CT done for restaging showed
no abnormal focus of increased tracer uptake in the scan as seen in MIP (a) and fused
transaxial images (b and c); triphasic diagnostic CT scan done a week after shows
subcentimeter size enhancing liver lesions in segment VIII and III (d and e), best
appreciated in portovenous phase. Thus, PET/CT did not detect the subcentimeter size
metastatic liver lesions. FDG: Fluorodeoxyglucose; PET/CT: Positron emission tomography/computed
tomography; USG: Ultrasonography; FNAC: Fine-needle aspiration cytology
SN, SP, PPV, and NPV of PET/CT was found to be 95%, 100%, 100%, and 75%, respectively,
with accuracy of 96% [[Table 4]].
Table 4: Result of Restaging
Discussion
Mucosal melanoma of anorectum is very aggressive disease with extremely poor prognosis
with overall 5-year survival rates of 3%–22%.[1],[2],[3] Most of the patients present late because of nonspecific symptoms. Furthermore,
majority of the lesions lack obvious pigmentation.[7],[8] Surgery with or without adjuvant radiotherapy is the mainstay of treatment for locoregional
disease. Despite aggressive local surgical control even in early-stage disease, local
and distant site recurrences occur frequently due to its multifocal nature and clinically
occult lymphatic spread.
PET, a whole-body metabolic imaging modality, appears ideal for accurate disease status
evaluation before undertaking curative resection. Malignant melanoma cells are extremely
FDG avid because of the upregulation of glucose transporter proteins, which forms
the basis of FDG PET imaging.
Combing anatomical imaging CT with metabolic imaging PET provides the added advantage
of anatomical localization and also increases the SN and SP by differentiating pathological
uptake from benign/physiological uptake.
Data about FDG PET/CT in anorectal malignant melanoma are sparse. Few case reports
and case series are noted in literature.[9],[10],[11],[12],[13] Studies in primary cutaneous malignant melanoma have shown the superiority of PET
and PET/CT over conventional imaging in initial staging and recurrent disease/restaging.[14],[15],[16],[17]
This study was performed with an aim to calculate the diagnostic performance of PET/CT
in initial staging and restaging with suspected recurrence in anorectal melanoma.
Patients for initial staging were treatment naïve and for recurrence had previously
undergone surgical resection, chemotherapy, or radiotherapy.
In the initial staging group, primary lesion was correctly identified in 57 patients
out of 61. In two patients, primary lesion was not evident on the scan of whom the
surgical specimen was positive. In both cases, the thickness of the primary lesion
was of about less than one centimeter size. This shows that small submucosal lesions
may be missed on FDG PET. Regional nodal metastases were correctly identified in 46
(75.5%) patients. PET/CT was true negative in 14 (23%) patients and false negative
in one patient. Postsurgery histopathology was positive for nodal metastases. Although
PET is the best modality to detect metastases in subcentimeter size nodes, no imaging
modality has yet proven accurate for the detection of micrometastases. This remains
a drawback of imaging. However, of all imaging modalities, FDG PET/CT has still the
best results.
Thus, in our study, PET/CT demonstrated high SN, SP, and accuracy for the detection
of primary tumor and locoregional nodes. The primary lesions and nodes which were
missed on PET/CT were of cm and subcentimeter size. It is known that PET may limit
the detection of lesions less than centimeter size due to finite spatial resolution
of the imaging system and due to partial volume effects.
The main strength of PET/CT, being a whole-body imaging, lies in the detection of
distant metastases. This was supported by our study, which showed a SN, SP, and diagnostic
accuracy of 100% for distant metastases.
The results of our study for the detection of regional nodes and distant metastases
in anorectal melanoma were similar to that of the results of case series by Falch
et al.,[9] who also found that PET/CT is superior for LN and distant metastases.
In our restaging cohort of 24 patients, PET/CT detected recurrence in 20 patients.
PET/CT was true negative in three patients and false negative in one patient in whom
liver metastases were missed.
The patient in whom PET/CT did not detect liver lesions was a FNAC-proven case of
metastatic liver lesion consistent with primary from anorectum prior to PET scan.
The patient was referred for PET/CT to rule out other sites of distant metastases.
No other lesions were found on PET/CT. Subsequently done diagnostic triple-phase CT
showed subcentimeter size enhancing liver lesions. The reason for false-negative results
of PET/CT may be due to subcentimeter size of the lesion and a contrast-enhanced low-dose
CT was done and not a diagnostic triple-phase CT scan.
Similar studies done in cutaneous melanoma for recurrent disease/restaging showed
that PET was superior for the detection of locoregional recurrence and distant metastases
with high SN and SP as compared to conventional imaging.[18],[19],[20]
Our study also revealed similar results with overall high SN, SP, PPV, and diagnostic
accuracy for the detection of recurrent disease. To our knowledge, this is the first
study in literature which looked at the role of FDG PET/CT in initial staging and
restaging of the mucosal melanomas of anorectum.
The limitations of our study include it was a retrospective analysis and the number
of patients included were small. However, considering the rarity of the disease and
lack of data on it, such studies are reported to strengthen the sparse existing literature.
Histopathological examination of all metastatic sites was not available as this is
not always feasible nor ethical in clinical practice.
Conclusion
PET/CT demonstrates a high diagnostic accuracy in the initial staging and detection
of recurrent disease in cases of anorectal melanoma. The aggressive nature and propensity
to metastasize needs a whole-body evaluation making FDG PET/CT the preferred modality
and one-stop imaging for anorectal melanomas.
