Imaging in Cutaneous Oncology: Radiology for Dermies




An overview of currently available imaging modalities is presented. Indications for imaging in cutaneous nonmelanoma skin cancers, selection of the most appropriate and cost-effective study, limitations, and risks are discussed. Finally, representative cases are discussed with emphasis on choice of imaging study for preoperative staging and treatment planning.


Diagnostic imaging has a supporting role in the evaluation and management of cutaneous neoplasms. As in radiology, the appearance of a lesion is a major factor in the process of reaching a diagnosis. When evaluating an aggressive skin tumor, cross-sectional imaging augments physical examination to refine the assessment of local spread, assist in determining regional nodal and distant metastatic involvement, and provide whole-body staging for cutaneous malignancies, such as melanoma, utilizing positron emission tomography-computed tomography (PET-CT).


The following issues should be kept in mind by the clinician when considering imaging for the evaluation of a suspected aggressive cutaneous neoplasm. First, the clinician must clearly communicate the pathologic diagnosis or suspected pathology and anticipated biologic behavior to the radiologist if a biopsy has not been performed. Secondly, the clinician should include specific references to particular concerns based upon the examination and patients’ symptoms: perineural invasion, depth of tumor, contiguous involvement of adjacent structures, and potential for local or distant metastases. Finally, the clinician should consult with the radiologist concerning the most appropriate and cost-effective examination to answer the clinical questions.


From the radiologist’s perspective, imaging has several general goals:



  • 1.

    Display anatomy desired


  • 2.

    Provide adequate detail


  • 3.

    Differentiate normal from abnormal structures


  • 4.

    Minimize risk to patients.



Within the broader field of surgical oncology, imaging also plays a role intraoperatively. For example, intraoperative magnetic resonance imaging (MRI) or ultrasound is employed during resection of some brain tumors; intraoperative cone beam CT, MRI, or ultrasound are employed during cryotherapy for solid neoplasms; and catheter angiography is utilized during solid-tumor embolization. Sentinel lymph node identification with vital dye and nuclear pharmaceuticals guides sampling intraoperatively. The accessibility of cutaneous neoplasms visually and the systematic mapping obtained during Mohs surgery likely in part account for the lesser role of intraoperative imaging for surgical management of the primary tumor. In the field of dermatologic oncology, imaging can certainly play a role in preoperative planning, discussion of potential morbidity with patients, preoperative staging, and post-therapeutic follow-up.


Imaging modalities


Plain Film


Most exhaustive dermatologic texts contain plain-film images of multiple odontogenic keratocysts, bifid ribs, and calcified falx cerebri seen in patients with nevoid basal cell carcinoma syndrome, subungual osteomas, or, unfortunately, enlargement of skull base foramina in patients with an aggressive cutaneous tumor. Medical imaging has advanced substantially since these images found their way into medical textbooks. The visual appearance and tactile characteristics of a cutaneous neoplasm often suggests the diagnosis prior to biopsy and sophisticated cross-sectional imaging serves to map out the extent of the abnormality or assist in staging if this is necessary prior to definitive intervention. As an example, evaluation of perineural tumor by plain film requires sufficient tumor spread and nerve expansion to cause smooth remodeling of bony foramina or gross erosion of adjacent bone caused by tumor infiltration. CT scanning improves sensitivity for bone changes, but still requires substantial soft-tissue abnormality to create visible secondary bone remodeling. Contrast-enhanced MRI provides greater soft-tissue resolution, and allows earlier detection of tumor infiltration along nerves or tissue planes prior to bulky nerve expansion or bone remodeling. Plain films are an interesting historical footnote but currently play no significant role in the workup of aggressive cutaneous malignancies, except perhaps with screening chest radiographs for metastatic disease.


Computed Tomography


Modern CT scanners employ a circular rotating x-ray tube head and detector array in combination with a continuously movable patient gantry to permit rapid scanning of a contiguous anatomic segment. Native collimation for multidetector CT scanners typically ranges from 0.5 mm to 1.0 mm and allows reconstruction of images with near isotropic resolution. This capability means that the smallest individual volumes (voxels) that make up the 3-dimensional data set are either only slightly rectangular or cubic at this 0.5-mm to 1.0-mm resolution allowing patient anatomy to be reviewed with reasonable fidelity in multiple planes. Image acquisition with modern equipment is quite rapid. For example, image acquisition time for the entire cervical, thoracic, and lumbar spine on a currently available 64-row scanner at 0.75-mm collimation is less than 30 seconds. The term spiral or more appropriately helical CT is applied to imaging performed in this manner. Tissue differences are displayed as a function of x-ray beam attenuation related to the predominant atomic number of the tissue or foreign material imaged. The relative attenuation of the x-ray beam is responsible for the areas of relative brightness or darkness on the image. Hounsfield units (HU) represent an arbitrarily defined scale of attenuation, with air at a value of -1000, fat at a value of -100, water at a value of 0, muscle at a value of approximately 40 to 50, acute blood at a value of approximately 70, and calcium and metal ranging from 200 to greater than 1000. Intrinsic differences between tissues often provide substantial resolution. For example, extraocular muscles and globes are clearly visualized against the lower density of the retrobulbar fat and inflammatory processes, which cause edema, and infiltration within subcutaneous fat is often visible without the addition of intravenous contrast.


Contrast is utilized to highlight normal and abnormal structures. For example, in the neck, contrast causes a typical enhancement pattern within major salivary glands, lymph nodes, and vascular structures, rendering this anatomy in greater detail than possible with an unenhanced scan. The thyroid also enhances after contrast, but has significant density caused by intrinsic iodine content and is usually easily visualized on an unenhanced scan.


Iodine linked to an organic molecule is utilized to provide contrast enhancement for CT. As described previously, x-ray beam attenuation is related to the atomic number of the material in the field of view. A typical dose of intravenous contrast (100 mL Ultravist 300) for a neck CT will cause attenuation reaching a value of approximately 180 to 200 HU within the vasculature, and will accumulate in tissues with impaired vascular barrier related to tumor, direct extravasation, or inflammation.


Commonly used iodinated contrast materials are iso-osmolar or slightly greater than typical whole blood osmolarity. High osmolar contrast agents are no longer widely utilized, and are associated with a higher incidence of nausea and vomiting. True allergy is uncommon, but mild reactions, including itching and hives, can be controlled or substantially blunted by the use of a combined oral premedication regimen, including prednisone and diphenhydramine. More serious reactions, including hypertension and airway compromise, should prompt strong consideration for withholding contrast. Renal insufficiency is primarily a concern with regards to precipitating acute renal injury in patients with borderline renal function. At one time, gadolinium-enhanced MRI was considered a viable alternative to the use of iodinated contrast in patients with borderline renal function. Subsequent experience with nephrogenic systemic fibrosis has obviously changed this approach. Consultation with a radiologist at your institution regarding optimal imaging for a given situation and protocols for managing contrast-related issues is advised.


The main advantages of computed tomography include speed, intermediate cost, easy accessibility at most institutions, and, in particular, fine bone detail permitting assessment for subtle sclerotic change or superficial bony erosion. Skull-base foraminal enlargement can be identified, but in the setting of suspected perineural spread indicates gross bulky disease. Benign entities, such as schwannoma or neurofibroma, may also enlarge bony foramina. In either case, MRI is superior to assess the skull-base foraminal contents or intracranial spread of disease. As previously mentioned, intrinsic attenuation differences between tissues with or without the addition of contrast provide some soft tissue resolution. Gross tumor extent, extracapsular nodal spread, central nodal necrosis, and broad survey for gross enlargement along draining lymph node groups can also be determined with contrast-enhanced CT.


CT has come under scrutiny as concerns over radiation exposure have risen. Many practices are now including the dose-length product, an estimate of radiation absorption for the body part and technical parameters in CT reports. Eventually, a total ionizing radiation exposure profile may be maintained as a part of a patient’s electronic medical record.


Ultrasound


Diagnostic ultrasound involves the use of a handheld transducer containing a piezoelectric crystal array, which generates sound waves within a narrow frequency spectrum. These sound waves travel through tissue and reflect back towards the transducer from areas of tissue interface. Mathematical reconstruction of this data creates a 2-dimensional, with some devices a 3-dimensional, image that can be viewed at any angle because of the free movement of the transducer held by the operator. Imaging is in real time and can include measurements of vascularity/quantitative blood flow because of the Doppler effect created by the moving blood through the vasculature included in the field of view. For the assessment of small or superficial lesions, a semisolid gel disk larger than the width of the transducer and approximately 1.0 to 1.5 cm in thickness (standoff pad) in conjunction with acoustically conductive gel is used to move the face of the transducer away from the surface of the skin.


The lack of ionizing radiation; real-time image reconstruction; small size of the transducer; and suitability for use with ancillary equipment, such as needle guides and cryogenic probes, makes ultrasound a key modality for intraprocedural imaging and, in particular, image-guided procedures. High-frequency, high-resolution ultrasound before Mohs surgery for basal cell carcinoma (BCC) and squamous cell carcinoma (SCC) has been studied and found to be suboptimal for identification of subtle areas of tumor extension, such as foci of dermal invasion for micronodular BCC and infiltrative SCC. Ultrasound is also useful for assessment of regional lymph node groups and is significantly more sensitive than CT alone to evaluate suspicious nodes and guide minimally invasive needle sampling. Experimental techniques, such as sonoelastography, show promise in identifying differentiating benign versus metastatic lymph nodes, but the offline processing is too time consuming to be useful in a busy clinical environment. Imaging depth limitations and the small size of the transducers make ultrasound less useful for wide survey examinations. Whole-body staging is primarily accomplished with PET or hybrid PET-CT scanners.


Nuclear Medicine


PET imaging utilizes 18-fluorodeoxyglucose (PET-FDG) to identify tissues with high metabolic activity. In particular, tumors with high glycolytic activity will accumulate 18-FDG, as the initial glycolytic metabolite becomes trapped within the cell after initial phosphorylation. PET scanning alone provides moderate-resolution cross-sectional information. Software can be employed to coregister PET data sets to a CT scan of the same patient’s anatomy, but is most effective in brain imaging because of the rigid outline of the skull and constraint of the intracranial contents. Hybrid PET-CT scanners improve upon this difficulty, with a PET detector ring merged on the same gantry with a multidetector CT scanner. Patients remain immobile during the scan, the 2 data sets are coregistered, and the low-dose CT scan obtained for coregistration also provides attenuation correction data for the PET study. The moderate-resolution metabolic information provided by the PET scan is displayed in an overlay with the higher-resolution CT scan for better anatomic correlation. Early work is underway on hybrid PET-MRI scanners, which would reduce the total ionizing radiation to patients and also benefit from the improved soft-tissue resolution obtainable with MRI.


PET staging sensitivity may be reduced for tumor volumes less than 1 cm and tumors with low metabolic activity. PET-FDG can detect metastases in areas of necrosis, scarring, and fibrosis secondary to radiotherapy. However, little data is available regarding staging for cutaneous nonmelanoma skin cancer. Fosko and colleagues reported on PET in 6 BCC greater than 1 cm of the head and neck, and in 3 subjects the PET imaging correlated with the size and extent of the soft-tissue invasion. All 3 of these tumors were of the nodular subtype. Two of the 3 tumors that did not highlight on PET imaging were infiltrative, and 1 was the nodular subtype. PET imaging did not detect perineural spread demonstrated by tissue biopsy. Beer and Weibel reported a recurrent BCC beneath a scar on the back from prior electrodesiccation and curettage that was detected at the time of a PET-CT performed for surveillance of other internal malignancies. Boswell and colleagues reported a subjects with metastatic BCC to the lung that was detected on PET-CT.


There are few reports of PET or PET-CT for evaluation of cutaneous SCC. Conrad and colleagues reported a subject with invasion of SCC into the pterygoid musculature and perineural SCC extending along V3 to the skull base confirmed on PET-CT. Cho and colleagues employed PET-FDG to subjects with stage 11 cutaneous SCC, 9 subjects had high-risk cutaneous SCC. A total of 25% of subjects had lymph node metastases, and 1 subject had lung metastases. All 9 high-risk SCC showed FDG uptake. Incidentally, 1 subject was also diagnosed with stomach cancer. Leach and colleagues reported 4 subjects with cutaneous SCC and 2 with melanoma who presented with cranial neuropathy as a sign of recurrent aggressive skin cancer, all 6 subjects exhibited symptoms associated with trigeminal nerve and 3 with facial nerve involvement. MRI confirmed perineural disease in all 6 subjects, and 1 subject had PET-CT imaging findings that correlated well with the MRI findings. The investigators emphasized the importance of early radiologic evaluation when patients present with signs of cranial neuropathy so that aggressive surgical and adjunctive therapy may be performed.


The clinical utility of PET/CT for head and neck tumors has been documented for initial staging and follow-up in rare head and neck tumors, evaluation of patients with an unknown primary, as well as follow-up for treatment response and surveillance for regional and distant metastases. Roh and colleagues reported on 24 subjects with rare head and neck tumors, 10 with melanoma, 9 with sarcoma, 3 with olfactory neuroblastomas, and 2 with BCC. PET/CT and CT/MRI scanning were performed at the initial staging or follow-up, and the diagnostic accuracy of CT and PET-FDG for detecting primary tumors and metastases were compared with histopathology. The PET-FDG and CT/MRI accuracies for detecting primary tumors were 92% and 79% respectively, and 91% and 74% respectively for nodal metastases. The sensitivity and specificity of PET-FDG for detecting distant metastases and second primary tumors were 100% and 87%, respectively. Rudmik and colleagues reported on 20 subjects with cervical metastases from an unknown head and neck primary. PET/CT increased the detection of a primary tumor from 25% with a standard work-up to 55%; there was 1 false negative PET/CT scan.


Kao and colleagues reported 80 subjects who received radiation therapy for head and neck cancer who were followed for a median of 21 months with clinical examination, PET/CT, and correlative imaging. The sensitivity, specificity, and positive and negative predictive values of PET/CT for detecting locoregional recurrence were 92%, 82%, 42%, and 98%, respectively; and distant metastases or second primary tumors were 93%, 96%, 81%, and 98%, respectively. Negative PET/CT results within 6 months after radiation therapy correlated with statistically significant improved 2-year overall survival rates. Gourin and colleagues reported 64 subjects with suspected recurrent head and neck SCC. Distant metastases were detected in 15 of 64 subjects, 13 of which were unsuspected prior to PET-CT. The sensitivity and specificity, and positive and negative predictive value of PET-CT in detecting distant metastases were 86%, 84%, 60%, and 95%, respectively.


MRI


Protons behave as small dipole magnets in the presence of the strong main magnetic field of an MRI scanner. A slightly higher proportion of the protons (hydrogen in organic molecules and water) align parallel to the main magnetic field along the z-axis of patients, which is the slightly lower energy state in comparison with antiparallel alignment, which is a slightly higher energy state. Within the static main magnetic field, these protons wobble or precess about the z-axis in an analogous fashion to a top that slowly wobbles about the vertical axis of gravity. The precession of the protons in the main magnetic field is rapid. At 1.5 T, which is the typical field strength of a clinical MRI scanner, protons precess about the z-axis at 63.7 MHz. This steady-state equilibrium is modified by applying a short-duration radiofrequency pulse, which causes these protons to be deflected from the main magnetic field axis, and also causes the protons to precess in phase with one another rather than the usual random precession. This brief change yields a weak but measurable signal, providing the data that is ultimately reconstructed to form an MR image. Intuitively, one might imagine a rapid realignment of the protons with the main magnetic field and a rapid return to random precession as soon as the radiofrequency (RF) pulse ends. Although this is a simplistic description, these 2 phenomena define T1 and T2 relaxation. The terms T1 or T2 are in essence time constants, which are unique for particular tissues or substances at particular magnet field strengths. Manipulation of the technical parameters of each individual scan yields the image characteristics specific for that particular pulse sequence. As one would expect, patient motion, abrupt changes in the local magnetic environment (for example, the dense bone of the petrous ridge and aeration of the mastoid air cells immediately adjacent to the ventral temporal lobes), and metallic or paramagnetic substances can sharply alter the data obtained and grossly distort the images.


In general, common substances responsible for a high signal on a T1-weighted scan include acute blood, highly proteinaceous material, fat, some valence states of calcium, and paramagnetic gadolinium contrast. A high signal on T2-weighted scans primarily reflect free fluid and tissues with high free water content, including areas of edema, tumors with moderate to high intracellular fluid; cerebrospinal fluid; cysts; vitreous and aqueous humor; and low-velocity intraluminal contents such as encountered in the hepatobiliary, gastrointestinal, and genitourinary tracts. Short tau inversion recovery (STIR) sequence suppresses all substances with short T1 relaxation, such as acute blood, fat, pus, certain valent states of calcium, and so forth. The purpose is to make tissues or collections with high free water content stand out intensely from a background of low signal.


Gadolinium-based contrast agents provide contrast by shortening T1 relaxation by a factor of approximately 680 yielding a high signal on a T1-weighted pulse sequence. Enhancement is increased in the presence of highly vascularized tissues, and in particular tissues with incomplete blood/tissue barrier as encountered in many tumors and in the presence of active inflammation. As previously stated, fat is also high signal on a T1-weighted pulse sequence. Application of an additional pulse sequence can be employed to null the high signal from fat and improve the visualization of normal structures or pathology exhibiting gadolinium enhancement. As an example, this reduces the intensity of fat signal surrounding the various neurovascular bundles exiting through skull base foramina, making these structures more conspicuous when there is abnormal enhancement from tumor infiltration, which is the purpose for postgadolinium fat suppressed T1-weighted scans commonly employed in the head and neck, body, and musculoskeletal MRI. Nemzek and colleagues reported a 95% sensitivity for MRI detection of perineural invasion, but only a 63% sensitivity for mapping the entire extent of perineural tumor.


For many years, gadolinium was thought to be a safe option for patients at risk for exacerbation of renal insufficiency from intravenous administration of iodinated CT contrast. More recently, nephrogenic systemic fibrosis (NSF) has come to attention and has sharply curtailed the use of gadolinium-based contrast agents in patients with renal insufficiency. A full review of NSF is beyond the scope of this discussion; the reader is referred to the recently published overview of NSF by Chen and colleagues. The key differences between the use of CT and MR contrast agents in patients with renal insufficiency revolve around the different mechanisms for possible adverse sequelae. Intravenous iodinated CT contrast administered to patients with marginal renal function can be associated with acute renal injury. The concern for NSF gadolinium-based contrast agents in patients with renal insufficiency is thought to be related to release and accumulation of free gadolinium in tissues. Normal or near normal estimated glomerular filtration rate (eGFR) is associated with significantly reduced risk for this complication. Adoption of restrictive policies regarding gadolinium-based contrast administration, and a switch from gadodiamide to gadobenate dimeglumine has reduced the number of NSF cases at the reporting institutions. Consult with the radiology department at your institution regarding the policies in place regarding administration of gadolinium-based contrast agents to patients with borderline renal function. At the authors’ institution, the cutoffs for mandatory postscan dialysis, nephrology consult and hydration, and proceeding with the examination are eGFR less than 15 mL/min, 15 to 30 mL/min, and greater than 30 mL/min, respectively.


Patient Characteristics and Imaging Modality


Radiologic studies may be indicated to assess the extent if the primary tumor, rule out metastatic spread, as well as follow up for local or regional recurrence or distant metastatic disease. The choice of imaging modality will not only be based on the suspected tumor biology and location but also on patient characteristics. Allergy to radiocontrast materials must be taken into account as previously discussed, as must renal function. MRI is contraindicated in patients with pacemakers and automated implanted cardiac defibrillators, ferromagnetic aneurysm clips, and spine stimulators. Other devices require modifications in scanning protocol. Implanted medication pumps must be turned off prior to scanning, interrogated for normal functioning, and turned on after the procedure. Vagal nerve stimulators may be scanned with modification to the typical pulse sequences employed. An exhaustive reference of implanted devices and relative safety at various field strengths is available in print and online at www.mrisafety.com .




Case presentations and discussion


Case 1: Recurrent Perineural BCC


A 53-year-old Caucasian woman with a history of Mohs surgery for a basal cell carcinoma of the right naso-facial sulcus 8 years prior presented with progressive elevation of the right ala and paresthesias of the right cheek ( Fig. 1 ). A punch biopsy adjacent to the linear scar revealed infiltrative basal cell carcinoma. Preoperative MRI revealed perineural tumor of the infraorbital nerve extending to the infraorbital foramen ( Fig. 2 ). A multidisciplinary approach with the facial plastic surgeon excising the tumor under general anesthesia and Mohs mapping of the excised tissue revealed clear peripheral margins, but perineural tumor extension to the infraorbital foramen ( Fig. 3 ). An anterior maxillotomy was performed, and V2 was traced posteriorly to the pterygopalatine fossa and foramen rotundum at skull base. The segment of V2 proximal to the infraorbital foramen was not involved with tumor, confirming the MRI findings. The patient underwent a complex reconstruction by the facial plastic surgeon with a right paramedian forehead flap, septal mucosal flap, auricular and septal cartilage grafts, and cervicofacial advancement flap with subsequent revisions of the right ala and a vascularized fat graft to the right cheek ( Fig. 4 ). She received postoperative radiation therapy, 6000 cGy in 30 fractions, and has no evidence of recurrence at 36 months postoperatively (see Fig. 4 ).




Fig. 1


A 53-year-old Caucasian woman with scar ( arrow ) from prior Mohs surgery at right naso-facial sulcus with right alar retraction.



Fig. 2


( A ) MRI with axial fat suppressed T2-weighted scan at level of nasolabial fold. Tumor mass ( arrow ). ( B ) Axial postgadolinium T1-weighted scan at same level showing enhancing tumor ( arrow ). ( C ) MRI with axial fat suppressed T2-weighted scan. Tumor showing increased signal and expansion superficial aspect of infraorbital nerve ( arrow ). ( D , E ) Axial and coronal postgadolinium T1-weighted scan showing enhancing tumor ( arrow ). ( F ) Coronal postgadolinium T1-weighted scan showing normal caliber infraorbital nerve 2 cm posterior to foramen ( arrow ). ( G , H ) Coronal postgadolinium T1-weighted scans at level of central skull base. ( G ) Normal orbital apex ( arrow ) and foramen rotundum ( arrowheads ). No tumor at posterior aspect of second division of right trigeminal nerve. ( H ) Normal third division of right trigeminal nerve ( arrowheads ), foramen ovale ( arrow ).



Fig. 3


Thickened infraorbital nerve extending to infraorbital foramen.



Fig. 4


One year after surgery.


The incidence of perineural BCC is 0.1% to 3.0%. Risk factors for perineural BCC include sclerosing, infiltrative and morpheaform growth patterns, recurrent BCC, and radiation treated tumors. Preoperative MRI was chosen for this patient in light of the clinical symptoms of perineural involvement of V2 and the recurrent nature of the BCC and the superior ability for MRI to detect the perineural tumor and the extent of soft-tissue invasion. Tumor appeared to extend to, but not beyond the infraorbital foramen. Knowledge of the potential extent of tumor spread centrally prompted the multidisciplinary surgical approach to include the Mohs surgeon, facial plastic surgeon, and skull base surgeon. Aggressive surgical resection of the involved nerve with clear histologic margins and adjunctive radiation therapy allow for the greatest chance of local control and lasting cure.


Case 2: Perineural SCC


A 62-year-old Caucasian man presented with an incompletely excised SCC after 2 wide excisions above the left eyebrow, with perineural invasion noted on the second excision specimen. Physical examination revealed a linear scar over the left eyebrow with paresis of the temporal branch of left facial nerve (noted after second excision, unclear as to whether this nerve deficit was from surgery or tumor involvement of the nerve)( Fig. 5 ). There was no palpable lymphadenopathy. Chest radiograph was normal. Five stages of Mohs surgery were performed, and the tumor extended medially from the left temple, almost exclusively along nerves, to involve the left supraorbital and supratrochlear nerve branches as well as the temporal branch of the left facial nerve, resulting in a 5.5 x 7.0-cm wound ( Fig. 6 ). MRI was performed and revealed no gross tumor surrounding the temporal branch of the facial nerve or the supraorbital nerve at the foramen ( Fig. 7 ). The wound was repaired with a combination of rotation flaps and the patient received postoperative radiation therapy. Three months following completion of radiation therapy, he developed a left-sided Bells palsy and left ear pain. MRI revealed thickening of the left facial nerve at the skull base extending through the stylomastoid foramen and into the vertical portion of the facial nerve canal ( Fig. 8 ). CT of the chest revealed a single bronchial metastasis, and a bone scan performed for right thigh pain revealed a single bone metastasis. The patient expired 18 months later despite chemotherapy and further radiation therapy to the lung and bone metastases.




Fig. 5


A 62-year-old Caucasian man with linear scar over left lateral brow at site of incompletely excised perineural SCC.



Fig. 6


Mohs surgery defect measuring 5.5 x 7.0 cm. Perineural SCC was noted around the temporal branch of the facial nerve, as well as the supraorbital and supratrochlear nerves, extending to the supraorbital foramen.



Fig. 7


( A , B ) MRI with pregadolinium and postgadolinium coronal T1-weighted scans showing scalp defect and some enhancing scar ( arrowheads ) at the margin of the resection. No gross tumor. ( C , D ) Pregadolinium and postgadolinium axial T1-weighted scans showing no tumor associated with the facial nerve in the left temporal bone ( asterisk ).



Fig. 8


( A , B ) MRI with postgadolinium coronal and axial T1-weighted scans. Reference line on coronal scan ( A ) is the axial plane in ( B ). Abnormal enhancement at the entrance to the stylomastoid foramen in the coronal plane and vertical portion of the facial nerve canal in the axial plane ( arrowheads ).


The incidence of perineural SCC is reported to be as high as 14%; risk factors include adenosquamous, spindle cell, and poorly differentiated subtypes. Only 30% or less of patients with perineural SCC or BCC will experience any neurologic sign or symptom associated with their neoplasm. Almost all patients with radiologic evidence of perineural invasion will be clinically symptomatic; the most commonly involved nerves are the trigeminal and facial nerves. Conversely, patients with clinical signs and symptoms of perineural tumor invasion may have negative MRI findings and should undergo surgical exploration. The rate of metastases to regional lymph nodes and distant sites is significantly higher with perineural head and neck SCC. Ballantyne and colleagues reported a case series of 80 subjects with head and neck SCC, 26 of which were primary cutaneous SCC. They noted that early major nerve trunk involvement could occur when invasive SCC developed near a nerve trunk, and that SCC invasion of smaller nerve fibers progressed axially to the central nervous system. Mendenhall and colleagues reported that perineural tumor spread can also extend peripherally.


Case 2 illustrates the need for aggressive initial surgery and radiologic work-up with MRI to detect perineural tumor, and that microscopic perineural tumor escapes detection by current MRI technology, as seen in the MRI performed immediately after Mohs surgery. CT and MRI have both been effectively used in the past to detect perineural SCC or BCC, prompting more aggressive treatment and resulting in improved outcomes. The development of Bell’s palsy in patients with facial SCC should prompt an immediate clinical and radiologic work-up and subsequent aggressive surgery with or without adjunctive radiation therapy in order to provide patients with the best possible long-term outcomes. In addition, in this patient the spread of SCC from the temporal branch of the facial nerve to branches of the first division of the trigeminal nerve illustrates the anatomic connections reported in cadaver studies by Li and colleagues. These investigators reported extensive anatomic connections in cadaver studies between the infraorbital nerve to the buccal branch of the facial nerve; the auriculotemporal nerve to the buccal, zygomatic, and temporal branches of the facial nerve; the supraorbital nerve to the zygomatic and temporal branches of the facial nerve; the mental nerve to the marginal mandibular nerve; and the buccinator nerve to the zygomatic, buccal, and marginal mandibular branches of the facial nerve. Clinicians should keep these anatomic connections in mind and perform both sensory and motor examinations on patients with tumors located near major nerve trunks on the face, as well as request radiologic surveillance of potentially involved sensory or motor nerves.


Case 3: Recurrent SCC of the Scalp


A 72-year-old Caucasian man presented with a fixed, tender 4-cm erythematous nodule on the left side of the vertex in the center of a surgical scar ( Fig. 9 ). This SCC had been widely excised twice and then treated with Mohs surgery and tissue expander-assisted rotation flap repair elsewhere for a SCC at this site 6 months prior. Review of outside operative reports revealed that the outer table had been burred down prior to the flap repair. There was no palpable lymphadenopathy. An MRI was performed, which revealed tumor extending through the skull to involve dura ( Figs. 10 and 11 A ). CT of the head was also performed showing tumor erosion through the calvarium and subtle thickening of the meninges suspicious for dural involvement (see Fig. 11 B–D). CT of the neck was negative. The patient underwent wide excision of skin, bone, and dura by a plastic surgeon and neurosurgeon. Intraoperative frozen sections and paraffin-embedded sections confirmed clear margins in the skin, bone, and dura specimens. The dural defect was repaired with a tensor fascia lata graft. Titanium mesh was secured and cranioplasty was completed with methyl methacrylate. The cutaneous 10 x 8-cm defect was repaired with a vascularized free flap from the anterolateral aspect of the left thigh. The patient received postoperative radiation therapy to the scalp and draining nodes. Two years later he presented with a left-sided neck mass that was biopsied by fine-needle aspiration and found to be metastatic SCC. He received radiation therapy to the neck and was without evidence of disease at 3 months follow-up.


Feb 12, 2018 | Posted by in Dermatology | Comments Off on Imaging in Cutaneous Oncology: Radiology for Dermies

Full access? Get Clinical Tree

Get Clinical Tree app for offline access