Top, Actinic keratosis on initial section; note the moderate cytologic atypia of keratinocytes confined to the lower half of the epidermis (asterisk). Bottom, Invasive squamous cell carcinoma on the 10th step section; note the presence of atypical cells and a mitotic figure (arrow) in the dermis.
Top, Actinic keratosis on initial section; note the moderate cytologic atypia of keratinocytes, budding of rete ridges (arrow), and elastotic dermis. Bottom, Basal cell carcinoma on the second step section; note the dermal nodule of atypical basaloid cells (asterisk).
Top, Actinic keratosis on initial section; note the moderate cytologic atypia of keratinocytes, budding of rete ridges (arrow), and elastotic dermis. Bottom, Seborrheic keratosis on the third step section; note the benign-appearing basaloid cells and horn pseudocyst (asterisk).
Carag HR, Prieto VG, Yballe LS, Shea CR. Utility of Step SectionsDemonstration of Additional Pathological Findings in Biopsy Samples Initially Diagnosed as Actinic Keratosis. Arch Dermatol. 2000;136(4):471-475. doi:10.1001/archderm.136.4.471
To discover additional diagnostic findings on step sections of biopsy samples showing features of actinic keratosis on the initial section and to correlate such findings with clinical and histological variables.
Prospective study comparing initial histological findings with those noted on deeper tissue levels.
University-based dermatopathology practice.
Fifty-seven patients (36 men and 21 women) with biopsy samples from 69 skin lesions.
Main Outcome Measures
Identification of additional pathological diagnoses in step sections and correlation with clinical diagnosis, size and location of lesion, history of skin cancer or immunosuppression, size and handling of specimen, and presence of ulceration on the initial level.
Additional diagnostic findings were present on step sections in 23 specimens (33%), including 9 (13%) with squamous cell carcinoma in situ, 3 (4%) with basal cell carcinoma, and 2 (3%) with invasive squamous cell carcinoma. Three variables were significantly correlated with the discovery of cancer on step sections: (1) ulceration on the first level, (2) clinical diagnosis of skin cancer, and (3) history of skin cancer diagnosed by biopsy examination. The latter 2 variables were also correlated with the discovery of any additional finding, whether benign or malignant, on step sections.
In biopsy samples initially diagnostic of actinic keratosis, examination of step sections contributes clinically important information. Step sections are particularly useful when a clinical diagnosis of skin cancer is present. The results of this study confirm the pathogenetic importance of actinic keratosis as a precursor to fully evolved malignant neoplasia and suggest that such lesions merit thorough histological study.
HISTOLOGICAL features of actinic keratosis (AK) are common in skin biopsy specimens taken from sites with long-term exposure to the sun and might be seen alone or associated with various benign and malignant lesions. Actinic keratosis is frequently the initial lesion in the sequence of tumor progression leading to invasive squamous cell carcinoma (SCC),1 and most cutaneous cases of SCC arise from AK.2 Moreover, AK as a marker of long-term solar irradiation also indicates indirectly an increased risk of basal cell carcinoma (BCC), although it is not necessarily an obligatory precursor lesion for BCC.3 These associations imply that a finding of AK on initial histological sections will lead to discovery of fully evolved carcinoma in some specimens if step sections are examined.
Despite the clear importance of diagnosing invasive or in situ skin cancer, few objective data exist to guide dermatopathologists in deciding whether to examine additional step sections in biopsy samples initially showing AK alone. Recent retrospective studies have demonstrated the utility of step sections for detecting cutaneous malignant tumors4 and for identifying stromal invasion in cases originally diagnosed as melanoma in situ.5 It has been suggested that evaluation of multiple sections of a "suspicious" AK might be necessary to rule out invasive SCC.6 However, there are no published data that help identify particular cases of AK in which deeper sections are most likely to provide additional diagnostic information. To address this need, we conducted a prospective study of specimens histologically diagnosed initially as AK to determine the frequency with which step sections yield additional diagnoses and to correlate such discovery with clinical and pathological variables.
Sixty-nine consecutive skin biopsy samples yielding an initial histopathologic diagnosis of AK were collected from August 24, 1998, to October 2, 1998. The specimens were obtained from 36 men and 21 women, aged 31 to 88 years (mean, 66 years), and represented approximately 6% of the total dermatopathologic specimens handled during the study period. All specimens were entirely submitted for histological processing after standardized prosection as follows: Shave biopsy samples measuring less than 5 mm and punch biopsy samples measuring less than 3 mm in diameter were submitted intact. Shave biopsy samples measuring 5 mm or more and punch biopsy samples measuring 3 to 5 mm in diameter were bisected. Punch biopsy samples measuring 6 mm or greater in diameter were divided at 2-mm intervals. Elliptical specimens less than 6 mm long were bisected longitudinally, and ellipses measuring 6 mm or more were serially sectioned transversely at 2-mm intervals. All slides were prepared from formalin-fixed, paraffin-embedded tissue cut 5 µm thick and stained with hematoxylin-eosin. For each specimen, the initial slide and all subsequent slides contained 1 to 2 ribbons consisting of 4 to 10 tissue sections each.
In each case, following our standard operating procedures, 10 step slides were prospectively cut at 50-µm intervals between slides and examined by the original attending dermatopathologist (V.G.P. or C.R.S.). Any additional findings (whether benign or malignant) and the level at which they were first detected were recorded. The 69 cases were then grouped according to the nature of the lesion, if any, discovered on step sections: benign lesion, malignant lesion, any additional diagnosis (including benign and malignant lesions), or no additional diagnosis.
Standard histopathologic criteria for diagnosis of AK and skin cancers were used, as follows: AK, moderate to focally severe cytologic atypia of keratinocytes, without replacement of the full thickness of the epidermis by atypical cells or with sparing of adnexal structures, plus the presence of solar elastosis in the dermis; SCC in situ, severe atypia of keratinocytes involving the full thickness of the epidermis at least focally but not extending into the dermis; invasive SCC, severe atypia of keratinocytes, breaking through the basement membrane zone and extending into the dermis; and BCC, atypical basaloid keratinocytes arranged in a peripheral palisade at the dermoepidermal junction and/or in the dermis and the presence of a myxoid stroma.
For all cases, clinic notes, operative reports, and gross specimen descriptions were reviewed for clinical diagnosis, size and location of the lesion, history of skin cancer, and current immunosuppressive therapy. Gross specimen size and handling (bisection, trisection, etc), and the presence or absence of ulceration on the original slide were recorded. These variables were then correlated with the additional diagnostic information gained from step sections using 1-way analysis of variance or χ2 tests.
Of 69 biopsy samples yielding an initial diagnosis of AK only, deeper sections revealed new diagnostic findings in 23 (33%) (Table 1). In 14 cases (20%), SCC (Figure 1) or BCC (Figure 2) was discovered on deeper sections. In total, the diagnosis was revised to SCC in 11 cases (16%), 2 of which showed tumor cells in the papillary dermis (ie, invasion). Step sections in 9 cases showed full-thickness atypia of the epidermis, without adnexal sparing (SCC in situ). However, intradermal involvement could not be completely excluded in 6 of these 9 lesions because the biopsy samples were superficial and little dermis was available for study; these cases therefore were diagnosed as "SCC, possibly in situ." The diagnosis of AK was revised to BCC in 3 cases (4%). In addition, various benign lesions (Table 1) were revealed on deeper sections in 9 (13%) of 69 cases; seborrheic keratosis was the most frequent benign lesion diagnosed from deeper sections (Figure 3). Only 2 lesions, diagnosed as invasive SCC on step sections, were reexcised and submitted to our laboratory for pathological examination; one reexcisional specimen had residual SCC and the other contained no tumor.
On average, benign lesions were discovered on the third deeper section and malignant lesions were discovered on the fifth; however, 2 cases of SCC required 9 deeper sections for detection. Of the clinical and pathological variables evaluated, 2 were correlated with the discovery of any additional diagnostic finding, whether benign or malignant: (1) a clinical diagnosis of a malignant lesion (P<.001) and (2) a history of skin cancer (P=.01). Three variables were significantly correlated with the discovery of malignant lesions per se: (1) clinical diagnosis of malignancy for the present specimen (P=.002), (2) presence of ulceration on the first histological slide (P=.01), and (3) history of skin cancer by biopsy examination (P=.03). The remaining variables (size and location of lesion, history of ongoing immunosuppression, and size and handling of specimen) were not significantly correlated with the discovery of additional diagnostic findings on step sections.
Deeper sections disclosed additional pathological findings in 33% of the cases in this prospective series of routinely processed specimens initially diagnosed as AK, a finding similar to the 37% rate reported in a recent retrospective study4 of step sections in dermatopathology in general. The high rate of discovery of new diagnostic findings affirms the potential benefit of examining deeper sections to obtain additional, clinically relevant information in biopsy samples initially diagnosed as AK.
Most important in our study was the discovery of skin cancers in 14 (20%) of these biopsy samples. The finding of fully evolved SCC in a proportion of cases with AK is not surprising in view of the widely held concept of AK as a "premalignant" lesion,1 a somewhat controversial term denoting a lesion intermediate in the multiple-step continuum that constitutes the process of carcinogenesis.7 Furthermore, the discovery of BCC in 3 cases (4%) reflects the similar pathogenesis (via chronic actinic injury) of AK and BCC and supports the concept of AK as a marker for increased skin cancer risk.
Several nonmalignant lesions were also detected on step sections in this series. Seborrheic keratosis (Figure 3) was an additional finding in 3 cases, perhaps simply representing collision of 2 common neoplasms8; however, sun exposure might play a pathogenetic role in the development of some seborrheic keratoses,9 and in exceptional instances SCC might arise directly in a seborrheic keratosis.10 Diagnostic features of porokeratosis became apparent on step sections in 2 cases. In one case, a fibrous papule was discovered in actinically damaged skin of the nose, probably representing the coincidence of 2 common lesions at this site. In another case, step sections revealed a proliferation of melanocytes in sun-damaged skin of the hand. These melanocytes lacked significant atypia and were interpreted as a likely consequence of long-term UV irradiation, again reflecting a common pathogenesis. The remaining 2 nonmalignant lesions were a comedo and a scar.
Economic concerns must be addressed in considering whether to perform exhaustive histopathologic examination of all cases yielding an initial diagnosis of AK. To do so would place an additional financial burden on the pathology practice because the supplementary costs of producing and examining step sections are not reimbursed by Medicare, managed care organizations, and most private insurance payers. The actual technical cost of producing the 10 additional slides in each case was $31.54 per block (direct cost=$14.71 and indirect cost=$16.83). Thus, the total additional technical cost for creating the 690 step sections in this series was $2176.26. By dividing this sum by the number of new diagnoses rendered as a consequence of examining step sections, one can estimate cost-effectiveness as a function of category of diagnosis achieved. Using this approach, there accrued overall additional technical costs of $94.62 per additional diagnosis attained (whether benign or malignant), of $155.45 per cancer detected (both in situ and invasive), and of $435.25 per invasive cancer detected. These figures are in line with those of several published cost-effectiveness studies in pathology. For comparison, one study11 concluded that clinically diagnosed cutaneous fibroepithelial polyps did not require microscopic examination because the global pathology charges to detect any malignancy or any potentially life-threatening malignancy in such specimens were $7430.61 and $37,153.05, respectively. In contrast, the same group suggested that it was cost-effective to examine clinically diagnosed seborrheic keratoses by histopathologic methods (at $937.55 to detect a malignant tumor).12 Netser et al13 concluded that it was cost-effective to examine slides of nonroutine tonsil and adenoid specimens because, on average, it cost $525 to render one consequential diagnosis in that set. In contrast, routine specimens had such a low diagnostic yield that it was considered cost inefficient (at $64,718 in charges per consequential diagnosis) to submit them for histopathologic study.
To improve the cost-effectiveness of examining step sections of AK, we sought to identify particular clinical or pathological features that might increase the likelihood of uncovering additional diagnostic findings. We found 3 such criteria that might help identify a subset of cases as candidates for deeper sectioning. The most significant factor correlated with the discovery of a malignant lesion on deeper sections was a clinical diagnosis of malignancy; also significant were a history of skin cancer and the presence of ulceration on the initial slide. Different strategies for case selection entail substantial trade-offs between sensitivity and specificity. If step sections had been obtained only in cases fulfilling all 3 of the previous criteria, 10 of 14 malignant lesions would have remained undetected, but only one case with no additional diagnosis would have been subjected to ultimately unfruitful step sectioning. On the other hand, if we had step sectioned all cases fulfilling 1 or more of the previous criteria, no malignant lesion would have been missed but 38 of 46 cases with no additional diagnosis would have been examined. If we had obtained step sections only from cases fulfilling 2 or more of the 3 criteria, 1 malignant lesion would have gone undetected but 21 cases with no additional diagnosis would have been examined.
With regard to the optimal number of deeper sections to obtain, a similar balance between sensitivity and specificity must be struck. Our data suggest that at least 5 deeper sections might be necessary to detect a significant number of hidden malignant lesions. However, if we had obtained 5 rather than 10 levels per case, 10 of 14 malignant lesions would have gone undetected. It is possible that examination of even more step sections through the entire tissue block would have led to detection of additional diagnostic findings. However, there is currently no generally accepted, standard method to direct the performance of step sections.14 Clearly, a difficult compromise must be reached between the desire for maximal diagnostic and prognostic information and the reality of scarce human and financial resources.
The presence of a clinical diagnosis of cancer was the factor most significantly correlated with discovery of malignancy on step sections (P=.002). This fact strongly argues that clinicians should provide thoughtful histories and diagnoses. The "black box" approach, in which pathologists are expected to arrive at the truth without access to relevant clinical data, carries a risk of misdiagnosis; lack of adequate clinical history is the most frequently reported deficiency in surgical pathology accessioning.15 Another highly significant marker for the presence of a malignancy on deeper sections was a history of skin cancer diagnosed pathologically (P=.03). These data show the potential benefit of performing an automated diagnostic search of pathology files on the patient, if available; this recommendation might not be practical for all laboratories, however.
The distinction between AK and SCC is sometimes subtle and, as our data show, may require careful step sectioning and expert evaluation. Our results therefore suggest that this is a diagnostic area in which subspecialty expertise in dermatopathology is desirable. Many managed care plans currently do not permit ready access to a dermatopathologist, particularly for "routine" diagnoses such as AK. This lack of availability of subspecialty dermatopathology experts represents a potential source of medicolegal liability for dermatologists who participate in managed care.
Because the designation AK might not adequately convey this lesion's potential for evolution to invasive SCC, the term "keratinocytic intraepithelial neoplasia" (KIN) was recently coined for a broad spectrum of lesions embracing AK and SCC in situ.16 Subclinical or flat lesions with focal atypia of keratinocytes of the lower one third of the epidermis would be designated KIN I. Hyperkeratotic papules with focal atypia of the lower two thirds of the epidermis would be designated KIN II. Plaques with diffuse atypia involving the full thickness of the epidermis would be designated KIN III. Using these criteria, the AK seen on our initial slides would correspond to KIN I or II and the SCC in situ cases would correspond to KIN III. The newly proposed KIN terminology has the virtue of reinforcing the concept of a morphologic and pathogenetic continuum between AK and SCC and might potentially facilitate needed reimbursement for treatment of AK.16 However, the diagnosis of KIN depends on correlation of gross and histological criteria, which may be difficult to achieve when, as in our practice, different individuals perform the clinical and pathological examinations. Moreover, the broad term KIN is potentially misleading, as it could in principle apply to a wider range of lesions (including seborrheic keratosis, Paget disease, and hidroacanthoma simplex), not just to the AK to SCC sequence; an analogous objection has been lodged against the term "vulvar intraepithelial neoplasia."17 Finally, it may be valuable for the time being to distinguish between AK and fully evolved SCC in situ (while recognizing their close relation) because their treatments might differ. For example, 2 commonly used treatments for AK—cryosurgery and topical fluorouracil therapy—had associated recurrence rates of 34% and 14%, respectively, when applied to SCC in situ.18 Conversely, complete surgical excision of all AKs is probably not necessary in view of their high rate of spontaneous disappearance.19
A commonly cited estimate of the rate of transformation from AK to frankly invasive SCC is 0.1% per lesion per year in untreated patients.19 However, because patients typically have numerous AK and are at risk for several years, the actual rate of malignant transformation for an individual patient might be on the order of 10% over a 10-year period.20 Lesions clinically diagnosed as hyperkeratotic AK might have an especially high rate of progression to SCC.21 Whereas relatively few AKs overall progress to fully evolved SCC, the converse frequency of AK in proximity to SCC ranges from 82%22 to 97%,23 indicating their common pathogenesis. Thus, the finding of AK has considerable clinical significance, justifying examination of multiple deeper sections in at least some patients at high risk for development of skin cancer.
In summary, when the initial sections reveal AK, it might be advisable to examine additional deeper sections to detect a fully evolved malignancy (BCC or SCC), especially when the lesion in question carries a clinical diagnosis of skin cancer. Other factors that also should prompt consideration of performing deeper sections include ulceration or a previous histological diagnosis of skin cancer in the patient.
Accepted for publication October 7, 1999.
This work was supported by the Department of Pathology, Duke University Medical Center, Durham, NC.
Presented in poster form at the joint meeting of the International Society of Dermatopathology, American Society of Dermatopathology, European Society of Dermatopathology, and Asian Society of Dermatopathology, New Orleans, La, March 17-18, 1999.
Reprints: Christopher R. Shea, MD, Department of Pathology, Duke University Medical Center, Campus Box 3712, Durham, NC 27710 (e-mail: firstname.lastname@example.org).