Clinical value of FDG-PET/CT in staging cutaneous squamous cell carcinoma

Abstract

Background:

Cutaneous squamous cell carcinoma (cSCC) is the second most common skin malignancy. CT/MRI are commonly used for staging however, the role of FDG-PET is not clearly established. In this study, we evaluated FDG-PET/CT imaging for initial staging of cSCC.

Methods:

FDG-PET/CT scans performed in patients with newly diagnosed cSCC were reviewed retrospectively. Images were visually assessed for lesions and FDG uptake (SUV) in primary and secondary sites was measured. Suspected lesions on FDG-PET/CT were correlated with histopathology when available, follow-up imaging or clinical data in others.

Results:

Twenty-three cSCC patients who underwent FDG-PET/CT at diagnosis were evaluated. Primary sites were in head/neck (n=21), chest (n=1), and foot (n=1). All patients had FDG-positive scans with total 51 FDG-positive lesions. All primary lesions (n=24) were FDG-positive (SUV2.3–22.8; mean 10.2), and additional 27 FDG-positive lesions, including 21 nodes, 4 cutaneous, 1 osseous and 1 lung lesion, were noted in 13 patients. Mean size of FDG positive nodes was 0.9cm (range 0.4–2.5cm), predominantly clinically impalpable. Pathology was available for 40/51 lesions; 31 sites positive for malignancy. SUV(mean±SD) was 9.2±6.2 for malignant and 2.7±1.2 for benign lesions. Sensitivity, PPV, and accuracy of FDG-PET/CT scan was 100%, 77.5%, and 77.5%, respectively. FDG detected 7 additional lesions in 3 patients, compared to CT/MR imaging. Overall, staging FDG-PET/CT detected 9 prior unknown lesions in 5 patients that were proven metastatic disease by histopathology or follow-up; FDG-PET/CT modified management in 5/23 patients (21.7%).

Conclusion:

FDG-PET/CT has high sensitivity in detection of cSCC lesions, including small cutaneous and nodal disease, and has potential role in initial staging and management.

INTRODUCTION

Cutaneous squamous cell carcinoma (cSCC) is the second most common skin malignancy, after basal cell carcinoma, with increasing incidence rates.^1–3 A recent analysis of Medicare databases from 2006–2012 state that approximately 5.4 million non-melanoma skin malignancies were diagnosed in 3.3 million patients and 38.2% of treatment procedures were performed for invasive cSCC.²

cSCC has several risk factors including cumulative UV radiation exposure, advanced age, immunosuppression, and fair skin; it is commonly seen in individuals with genetic predisposition or in people living in certain geographic locations.⁴ Clinical course depends on factors including size and location of primary tumor, stage at presentation, histological factors, lymph node involvement, and distant metastasis.⁵ cSCC mostly metastasizes to locoregional lymph nodes, with a nodal involvement rate of up to 10–25% in those with high-risk factors such as tumor invasion >6mm thickness, immunosuppression, primary site in ear, presence of desmoplastic growth, and poor differentiation/undifferentiated tumors.^4,6,7 Distant metastases occur in less than 5% of cSCC patients.⁸ Involvement of loco-regional nodes and distant disease affect overall prognosis and survival.⁹ A higher rate of recurrence is seen in patients with nodal and distant metastases.

Radiological studies including computed tomography (CT) and/or magnetic resonance imaging (MRI) are generally performed for staging, preoperative assessment of local and deep soft tissue extension, osseous involvement, perineural invasion, and lymphatic spread.¹⁰ MRI is preferred for evaluation of perineural involvement and intracranial extension.¹¹ In general, CT and MRI are performed for smaller regions, limiting assessment of lesions outside the field of imaging. Additionally, assessment of active disease in small lesions and nodes is limited due to size criteria.^12–14

cSCC cells express GLUT-1 and have the ability to accumulate [18F]2-Fludeoxyglucose (FDG).^15,16 FDG-PET/CT has high sensitivity in detection of cutaneous lesions as demonstrated in other skin malignancies.^17,18 Its role in routine evaluation and staging of cSCC has not been clearly established. NCCN guidelines recommend the use of FDG-PET/CT, at the clinician’s discretion, to rule out distant disease in patients with locoregional nodal disease and for possible use in RT planning.¹⁰ For nodal involvement, studies report sensitivity of 91–100% and specificity of 81–90% on staging FDG-PET/CT in cSCC patients.^19,20 A few small studies have shown potential of FDG-PET/CT to modify management in 6.2–40% of cSCC patients.^20–22 The common limiting factor in many of these studies was a mixed patient selection criterion with use of FDG-PET/CT at different clinical phases and treatment. Limited data exists for use of FDG for primary staging; mainly confined to case reports and small case series. In this retrospective analysis, we analyzed the diagnostic performance of FDG-PET/CT in staging patients with newly diagnosed cSCC.

MATERIALS AND METHODS

Study design and patient population

We performed a retrospective analysis of patients with biopsy-proven cSCC who underwent FDG scan at diagnosis in our institution from 2000–2016. The study was approved by the Institutional Review Board in compliance with the Health Insurance Portability and Accountability Act regulations. Patients with newly diagnosed cSCC of any region were included, except mucosal or ano-genital SCC or with another active malignancy. All FDG scans were reviewed independently. Electronic medical records were reviewed for clinical, laboratory, and pathologic data.

FDG-PET/CT scans

FDG-PET/CT scans were performed per routine clinical practice. Following at least 6 hours of fasting and blood glucose level ≤200 mg/dl, patients were injected with 370–555MBq of FDG. The scans were acquired 60–90minutes post-injection from base of skull to mid-thigh in 12, vertex to mid-thigh in 5, and vertex to toes in 6 patients. All scans included a low-dose CT for attenuation correction. Images were reviewed on PACS workstation (AW suite 2.0; GE Healthcare) displaying maximum intensity projection and multiplanar PET, CT, and PET/CT fusion images. Standard uptake value (SUV) normalized to body weight was measured for lesions.

Image analysis

FDG-PET/CT scans were reviewed by two independent nuclear medicine physicians. All focal areas of increased FDG uptake clearly above the normal background and not associated with physiologic sites of uptake were considered suspicious sites/lesions and noted. Size and maximum standardized uptake value (SUV) of lesions were measured.²³ All lesions were correlated with pathological findings where available, and in the rest, findings were correlated with follow-up imaging (up to at least 6 months) or clinical data.

Pathological correlation

FDG-avid suspicious lesions, corresponding to primary or secondary sites, were correlated with surgical histopathology or biopsy, wherever available.

Statistical analysis

FDG-positive lesions confirmed by pathology were classified as true positive. Sensitivity and specificity of FDG-PET/CT was calculated based on pathologic correlation. Medians and ranges were calculated to summarize continuous variables.

Change in management

The original treatment plan based on clinical assessment and conventional imaging was compared to the final treatment plan implemented after additional FDG-PET/CT staging to evaluate the impact of FDG-PET/CT on patient management.

RESULTS

Twenty-three consecutive patients (median age 76y; range: 54–101y) who underwent FDG scan at initial staging were analyzed. Patients were clinically staged based on American Joint Committee on Cancer (7^th edition²⁴) criteria. Details on patient and tumor characteristics are summarized in Table 1.

Table 1.

Patient and tumor characteristics

Number of patients	23
Median age at diagnosis, years (range)	76 (54–101)
Median follow up time, months (range)	10.8 (2.8–93.8)
Sex, n (%)
Male	20 (87%)
Female	3 (13%)
Number of tumors
Ear and lip	5 (21.7%)
Head and neck (except ear and lip)	16 (69.6%)
Trunk	1 (4.3%)
Foot	1 (4.3%)
Tumor diameter (radiological)
<2.0 cm	8
> or = 2.0 cm	15
Histologic differentiation
Well and moderately	8 (34.8%)
Poor	15 (65.2%)
Perineural invasion
Absent	9 (39.1%)
Present	8 (34.8%)
Not known	6 (26.1%)
Primary treatment
Surgery	7 (30.4%)
Radiation	5 (21.7%)
Surgery + Adjuvant Radiation	8 (34.8%)
Chemoradiation	3 (13.1%)
AJCC* tumor stage
I	0
II	11 (47.8%)
III	7 (30.4%)
IV	5 (21.7%)

^*AJCC: American Joint Committee on Cancer (7^th edition)

Findings on FDG-PET/CT

All patients (n=23) had positive FDG scans (Fig.1) with FDG-avid primary lesions. Ten patients had FDG-positive disease only in the primary site and 13 patients had FDG-avid additional sites including locoregional node(s) (9 pts), skin lesion (1 pt), both skin lesion and regional node (1 pt), both bone lesion and regional node (1 pt), and lung nodule (1 pt).

Fig. 1.

90-year-old man with primary cSCC in scalp was referred for staging FDG-PET/CT scan that demonstrated FDG-avid 2 cm primary lesion; SUV: 2.3 A, Maximum intensity projection, long black arrow B, PET coronal, black arrow C-D, Fused PET/CT coronal and axial, white arrow; and FDG-avid left parotid node on fused PET/CT imaging E, PET/CT axial F, CT axial with SUV of 13.7, measuring 3.5 x 2.6 cm. The patient underwent surgery and received adjuvant radiation to the site of the primary and left parotid gland.

A total of 51 lesions were seen on all FDG scans: 24 primary and 27 secondary sites. Primary sites included 5 scalp lesions (4 pts), 4 temple (4 pts), 4 ear canal (4 pts), 1 lip (1 pt), 4 cheek (4 pts), 1 supraorbital (1 pt), 1 lower eyelid (1 pt), 1 left nasal (1 pt), 1 neck (1 pt), 1 chest (1 pt), and 1 foot (1 pt). One patient had 2 primary cSCC lesions on the scalp.

Mean size for primary lesion was 2.8cm (range: 1.3–11.2cm) with 7/24 primary lesions (30%) measuring < 2cm (Fig.1). The smallest and largest lesions measured 1.3cm and 11.2cm (foot), respectively. FDG-positive secondary sites (n=27) included nodes (n=21), skin lesions (n=4), lung (n=1), and bones (n=1). Mean size of FDG-positive nodes was 0.9cm (range:0.4–2.5cm), with 17/21 nodes <1cm and one node <0.5cm (Fig.2). Of 21 nodes detected, only one parotid node measuring 2.6cm in short axis and another node in level II cervical region measuring 0.8cm in short axis were clinically palpable.

Fig. 2.

83-year-old man with primary cSCC in right temple underwent staging FDG-PET/CT scan that demonstrated increased uptake in primary site in right temporal scalp A, MIP image of head and neck; SUV: 21.5. Additional FDG-avid sites included a right parotid node B, PET/CT fusion axial (SUV 9.0); and right level II cervical lymph node measuring 1.5 x 0.8 cm C, PET/CT fusion axial (SUV 7.8). Subcentimeter right supraclavicular lymph node D, PET/CT fusion axial (SUV: 3.5) (white arrow). The patient received chemotherapy and radiation therapy to neck for nodal disease.

Mean SUV for primary lesions was 10.2 (range:2.3–22.8). The lowest uptake was noted in a primary cheek lesion measuring 1.3cm (SUV 2.3), while the highest uptake was seen in a lesion in supraorbital region measuring 4.0cm with SUV 22.8 (Fig.3). FDG-avid lung nodule, SUV 1.3, was detected in one patient, while nodal lesions showed an average SUV of 3.9. Four FDG-positive cutaneous sites showed mean uptake of 6.4 (range: 2.7–13.0).

Fig. 3.

84-year-old man with primary cSCC in right supraorbital region was referred for staging FDG-PET/CT scan that demonstrated FDG-avid primary lesion A, MIP lateral image B, MIP anterior image C, Fused PET/CT axial (SUV: 22.8); and focal FDG uptake in right intra-parotid node (SUV: 4.2), measuring 1.1 x 0.7 cm D, Fused PET/CT axial (white arrow) and E, CT axial (white arrow). The patient was treated with radiation therapy for both primary site and FDG-avid node.

Conventional imaging(CI) including CT/MRI of head and neck, performed within 4 weeks of PET, was available for comparison in 12 patients. FDG-PET/CT showed additional lesions in 3 of these 12 patients with a total of 7 additional lesions. Abnormal nodes were detected in 5/12 on CT/MRI, compared to 7/12 patients with FDG-PET. FDG-positive additional lesions included 6 neck nodes that were not palpable clinically (in 2 patients) and 1 nasal skin lesion (in 1 patient). Three of the six nodes were metastatic based on follow-up, while the skin lesion was positive for basal cell carcinoma (BCC). Beyond the field of CT imaging, FDG-PET/CT showed additional disease in two patients including one cutaneous lesion on arm that was positive for BCC and one preauricular node positive for metastatic cSCC on follow-up. No lesions were noted that were positive on CI but PET-negative.

Pathologic correlation

Of the total 51 FDG-positive sites, pathological correlation was available for 40 lesions including all 24 cSCC primary and 16/27 secondary sites of uptake. Fifteen patients underwent surgery for primary tumor including 7 wide-local excisions and 8 radical resections, allowing for surgical histopathology correlation for 16 primary lesions. In the remaining 8/23 patients, biopsy correlation was available for 8 lesions.

For 16/27 secondary sites in 9 patients, neck dissection ± parotidectomy was performed in 5 patients, which allowed for histopathologic correlation of 6 nodes while biopsy correlation was available for remaining 10 lesions including 6 nodes, 1 lung lesion, 1 bone lesion, and 2 skin lesions. Three of the total 12 nodes were positive for metastatic cSCC on pathology, while the other 9/12 nodes were benign (2/9 showing reactive lymphoid hyperplasia and 3/9 showing non-necrotizing granulomatous inflammation). One bone lesion was also proven to be cSCC on biopsy. The lung lesion revealed squamous cell carcinoma (SCC); however, metastatic versus new primary etiology was unclear on pathology. Two FDG-positive secondary skin sites revealed BCC on pathology.

Statistical analysis

Using pathology correlation, the sensitivity, positive predictive value (PPV), and accuracy of FDG-PET/CT scan was 100%, 77.5%, and 77.5%, respectively (Table 2).

Table 2.

Probability table to calculate sensitivity and PPV and accuracy.

	Sensitivity	PPV	Accuracy
Primary disease	100%	100%	100%
Nodal metastases	100%	25%	25%

Analysis with follow-up imaging

Of the remaining 11/51 FDG-positive lesions that did not have pathologic correlation, follow-up imaging including CT or FDG was used for assessment, allowing for confirmation of 9/11 lesions. Six locoregional nodes in the head and neck region, that were not palpable clinically, including cervical, submandibular, and pre-auricular nodes were considered suspicious based on clinical judgment and were irradiated. Follow-up imaging after 6 months post-treatment showed no evidence of disease in these sites.

Three FDG-avid neck nodes, only one palpable clinically, seen in a patient with primary in right temple showed an increase in size and metabolic activity on follow-up FDG-PET/CT performed two months later, without interim treatment, and were treated with radiation. Two FDG-positive skin lesions could not be confirmed due to lack of follow-up.

Change in management

Overall, FDG-PET/CT had a positive impact on management in 5 patients (21.7%, 5/23) and led to modification of the treatment. Two patients with primary in lip and ear received radiation dose modification based on FDG-positive nodes located in submandibular and cervical level II region that were not palpable clinically. In one patient with primary supraorbital lesion, radiation field was extended to include an FDG-positive pre-auricular node, also not palpable clinically. One patient with FDG-positive lung nodule underwent excision with video-assisted thoracic surgery and was confirmed SCC. Another patient with BCC lesions received radiation to arm and cheek.

DISCUSSION

cSCC is frequently detected in early stages and can be treated with surgery, radiation, topical therapy, or photodynamic therapy. Patients with high-risk disease are at increased relative risk of recurrence, nodal involvement, distant metastasis, and disease-specific death rate.²⁵ Accurate identification of disease in such patients is critical for deciding optimal treatment plan. Evaluation of small cutaneous lesions is limited on CT/MRI and interpretation of nodal disease, specifically in nodes that are not palpable clinically and are smaller than pathological size criteria on conventional imaging, may be another limitation. Our results showed prominent uptake in all primary lesions with 100% sensitivity, predominantly located in head & neck region. The nodal metastasis was 26.1% (6/23 patients) based on pathology and follow-up imaging; nodal stations involved included level II cervical, preauricular, parotid and supraclavicular. The nodal metastases were noted in patients with either stage III or IV disease with variable sizes of the primary lesion ranging from 0.9 – 4.0cm. None of our patients showed distant metastasis. No specific pattern was seen with the stage at presentation, histology of primary lesions, nodal and distant metastases, possibly limited assessment due to small patient numbers. In our study, FDG-PET/CT detected nodes that were clinically impalpable and while the data is small, FDG-PET/CT may be a valuable tool for staging in patients not only to rule out distant metastases but also for pre-operative nodal assessment and staging. More data and dedicated studies are needed to further evaluate this.

Overall, staging FDG-PET/CT detected 9 prior unknown lesions in 5 patients that were proven metastatic disease by histopathology or follow-up; these would have been otherwise missed in the absence of FDG-PET/CT imaging. Additionally, we found that FDG-PET/CT impacted a change in management in 5/23 patients (21.7%), including detection of second primary in two patients; basal cell carcinoma and synchronous lung malignancy. These results are concordant with other studies that have reported nodal involvement sensitivities ranging between 91–100% and change in management in 6.25–40% of patients.^19–21 A recent study in node-positive cSCC patients calculated the positive and negative predictive value for nodal detection in preoperative FDG-PET/CT as 91.1% and 66.7%, respectively.²⁰

In comparison to conventional imaging (CI), FDG-PET/CT offered advantage in disease detection in small nodes that were negative by size criteria on CT. These nodes, considered suspicious, were irradiated and showed resolution in follow-up imaging. Small cutaneous lesions can also be difficult to detect and characterize on CI. FDG-PET/CT allows for whole-body survey, enabling detection of metastatic sites as well as secondary malignancies. In our study, FDG allowed for detection of additional skin lesions in two patients (8.7%), one had two biopsy-proven BCC treated with radiation to cheek and arm (Fig.4). Additional incremental value of FDG-PET/CT is the possibility of detecting metastatic lesions outside the field-of-imaging of CT/MR, which are conventionally acquired as single or limited field-of-view. In our study, one patient showed FDG-positive lung lesion that was SCC.

Fig. 4.

70-year-old man with primary cSCC in scalp underwent staging FDG-PET/CT scan that demonstrated additional skin lesions A, MIP image, dotted arrow; and B, PET/CT fusion axial in cheek (SUV: 6.3) A, MIP image, solid arrow; and C, PET/CT fusion axial in left arm (SUV: 3.5). These were basal cell carcinoma, nodular subtype on pathology and were treated with radiation therapy.

In contrast to prior reported studies, our analysis is focused on patients with initial diagnosis and staging with availability of histopathological correlation for most of the lesions, while others have primarily reported data on mixed clinical populations. Additionally, the patient selection and use of FDG-PET/CT and radiographic evaluations in other studies were not standardized for assessment. The number of patients assessed with FDG in these studies is also small. Ruiz et al evaluated the impact of radiological imaging in 98 high-stage cSCC patients; however, the predominant imaging modality used was CT and only 10 patients received FDG-PET/CT.²²

The use of FDG-PET/CT can lead to an incremental value for assessment of staging cSCC. The limitations, however, include the detection of false positives. In our analysis, false positives were noted in 5 patients, mainly involving the nodes. These were in the neck, inguinal, and pelvic region and SUVs were generally low (range:1.3–5.4), though overlap is seen with high SUV of 5.4 in an inguinal node. We noted a false-positive rate of 17.6% (9/51 lesions), which is comparable to results seen in other studies reporting false-positive rate of up to 16.6% (range: 1.5–16.6%).^18–20 Five of our patients with primary in scalp (n=2), ear (n=2), and foot (n=1) had 9 false-positive locoregional nodes in the cervical, occipital, preauricular, parotid, inguinal, and external iliac regions. All these nodes except the inguinal were subcentimeter-sized and showed low-grade, nonspecific FDG uptake (mean SUV: 2.1, range: 1.3–2.9). The inflammation of primary sites likely led to involvement of the draining lymph node basin, thus explaining FDG accumulation. The patient with the primary foot cSCC had an associated large ulcer with overlying granulation tissue that was probably the cause of FDG accumulation in ipsilateral inguinal and external iliac nodes. Also, pathology of cervical nodes from the patient with the primary ear cSCC showed non-necrotizing granulomatous inflammation. A prior study by Fujiwara et al used hematoxylin-eosin stained preparations to demonstrate increased presence of inflammatory mononuclear cells at the cSCC primary site in false-positive scans relative to the true-positive or true-negative group; sensitivity and specificity of FDG-PET/CT for the detection of LN metastases was 100% and 81.5%, respectively.¹⁹

False-negatives were not seen in our study, though prior studies have reported up to 13–25% of false-negatives. Cho et al found false-negative lesions in 3 of 12 patients (primary site in forearm, vulva, and cheek) and Supriya et al noted false-negative neck nodes in 4/31 patients.^16,21

Another advantage of whole-body FDG-PET/CT is that it may detect secondary skin malignancies in cSCC patients since they are at increased risk.²⁶ In our study, two nodular subtype BCC lesions were found in one patient. Similar findings were seen in the retrospective study by Cho et al where FDG accumulation was noted in 3/6 BCC lesions.^27,28 It is known that overexpression of GLUT-1 receptors in BCC cells, associated with areas of squamous metaplasia, leads to increased FDG uptake.¹⁵ Given the possibility of improved disease detection, incorporation of FDG-PET/CT into initial staging evaluation should be considered.

Our study is limited due to its retrospective design and small sample size. As FDG-PET/CT is not integrated in routine evaluation at staging in all patients, we were limited in number. Additionally, cross-sectional imaging for a head-to-head comparison with FDG-PET/CT scan was only available for some patients that limits assessment.

CONCLUSIONS

FDG-PET/CT can identify primary and locoregional disease in cSCC patients with high sensitivity, including small cutaneous and clinically impalpable nodal lesions, and thus has a potential role in initial staging and modifying subsequent treatment strategy. Further assessment in larger number of patients is needed to establish its role in routine clinical use.

Abbreviations used:

FDG-PET/CT

[18F] 2-Fludeoxuglucose positron emission tomography/computed tomography

cSCC

cutaneous squamous cell carcinoma

SCC

squamous cell carcinoma

BCC

basal cell carcinoma

CT

computed tomography

MRI

magnetic resonance imaging

GLUT-1

Glucose transporter 1

UV

ultraviolet

HIPAA

Health Insurance Portability Accountability Act

MIP

maximum intensity projection

PACS

picture archiving and communication system

AJCC

American Joint Committee on Cancer

TP

true positive

PPV

positive predictive value

SUV

standard uptake value