Painless Photodynamic Therapy for Facial Actinic Keratoses: A Retrospective Cohort Study of the Post-Treatment Inflammatory Response

Abstract

Introduction:

Photodynamic therapy (PDT) is a safe, non-mutagenic, and non-scarring treatment for actinic keratoses (AK).

Background:

‘Painless’ photodynamic therapy (p-PDT) is a regimen for AK that employs simultaneous aminolevulinate incubation and blue light illumination. The efficacy of p-PDT resembles that of traditional PDT, but detailed mechanisms of action for p-PDT are not well understood.

Methods:

To characterize the inflammatory effects of the p-PDT procedure 48 hours following treatment and determine the association of inflammation with precancer burden, we performed a retrospective cohort study of 104 patients with AK of face or scalp treated with p-PDT between 2017 and 2019. Patients self-reported their side effects 48 hours following p-PDT and took photographs of their face and scalp. Photographs were edited to define seven anatomic regions, and erythema was scored by four investigators.

Results:

Ninety-eight patients provided photographs suitable for erythema evaluation. Most patients experienced 2 or more side effects and some pain 48 hours post-procedure. Females experienced more pain (p = 0.01) and side effects (p = 0.002) compared to males. AK burden was positively associated with post p-PDT erythema response (p < 0.0001) at all sites, but particularly in the temples (p = 0.002) and supralabial area (p = 0.009).

Discussion:

This study confirms a strong clinical inflammatory response after p-PDT. Severity of inflammation is positively associated with AK tumor burden, suggesting that post-treatment inflammation may be a pre-requisite for p-PDT efficacy. Interestingly, the results also identify certain gender-related differences in the severity of side effects experienced by patients post-PDT.

INTRODUCTION

Actinic keratoses (AK) are premalignant squamous lesions of the skin that occur in chronically sun-damaged areas, and are the result of field cancerization ¹. AK can be well-managed using photodynamic therapy (PDT) in which the prodrug 5-aminolevulinic acid (ALA) is applied to precancerous skin cells that selectively absorb ALA and convert it into protoporphyrin IX (PpIX) which is subsequently activated by either blue or red light ². This causes a photoreaction that produces reactive oxygen species (ROS), induces apoptosis, autophagy, and immunogenic cell death, resulting in eventual lesion clearance ^{3, 4}. Although PDT has many benefits, including its ability to target multiple lesions without causing scarring and hyperpigmentation, patients frequently experience unpleasant side effects (e.g., pain and burning) during the treatment which many patients find undesirable, leading to underutilization of the technique ^{5, 6}.

To reduce the pain associated with PDT, studies aimed at providing a “painless” PDT experience were developed, such as a 2016 study demonstrating effective lesion clearance and minimal pain with 15-minute ALA exposure compared to conventional PDT⁷. In 2020, our group described a new PDT regimen that is nearly painless during illumination, yet still provides excellent lesion clearance ⁸. The regimen has been replicated by others, with similar results ^9,10. The aspect of the painless PDT treatment (p-PDT) which we reported that differs from previous traditional regimens is that illumination is begun immediately after ALA application, without any extended ALA incubation. Interestingly, the mechanism of action for p-PDT has become a subject of interest because it differs markedly from that of conventional (long-incubation) PDT. In preclinical studies in hairless mice, we showed that unlike with conventional PDT, there is minimal cell death and minimal generation of reactive oxygen species (ROS) within AK lesions in the first 24 hours after p-PDT ⁴. Instead, a delayed inflammatory response is seen, beginning several days after treatment, that appears to be largely responsible for the eventual clearance of lesions because AK cell death was not observed immediately after p-PDT ⁴. An unanswered question was whether or not such an inflammatory response might also occur after p-PDT treatment of human AK.

A strong clinical inflammatory reaction has always been a prominent feature of conventional PDT¹¹. Conversely, inflammation after p-PDT shows a delayed onset, but the full extent of the inflammatory response after p-PDT remained to be characterized. Accordingly, we set out to determine the extent of the clinical inflammatory response in patients following p-PDT. This was done in a real-practice setting, with a large patient sample size. Patients with AKs of the face or scalp, undergoing routine painless PDT in our clinics, were surveyed regarding their experience with inflammation-related side effects and usage of topical steroids for relief in the first 48 hours after p-PDT treatment. Patients provided photographic documentation of erythema at 48 hours post-PDT. The study evaluated the following hypotheses: (1) p-PDT generates a significant inflammatory reaction as measured by visible erythema, and (2) the intensity and extent of inflammation are associated with AK disease burden and anatomic location. The potential significance of this study included the possibility of obtaining additional insights regarding clinical outcomes after p-PDT; mechanisms of action of p-PDT; and associations between the inflammatory response and patient-specific characteristics such as neoplastic disease burden.

METHODS:

Study Design

This retrospective study was conducted at a single academic institution, using a database registry in which patient responses to a post-PDT electronic aftercare visit (E-visit), including self-generated patient photographs, had been collected. At routine PDT treatment visits, patients were invited to respond to an E-visit invitation which contained an online survey asking about side effects at 48 hours after p-PDT treatment. They were also invited to upload photographs of their face and scalp using a smartphone and good lighting; these were stored in their electronic medical record. Photographs were then edited by one of our investigators, who adjusted the brightness and contrast so that the periorbital (unexposed) skin looked as normal as possible, and also drew dotted lines to define the boundaries of 7 anatomic areas of interest (forehead, temples, cheeks, ears, supralabial, chin, and scalp), as shown in Supplemental Figure 1. Edited photographs were utilized for erythema scoring (see below). The study was approved by the Cleveland Clinic Institutional Review Board (IRB #22–447, approval date 5/16/2022).

Study Population

The study consisted of 104 patients with actinic keratoses of the face or scalp who were treated with painless PDT between 2017 – 2019, and completed an aftercare E-visit with at least one good quality photograph.

Interventions

All patients were treated with the recently described painless PDT regimen⁸. Briefly, ALA 20% solution (Levulan Kerastick; Sun/DUSA pharmaceuticals, Wilmington, MA) was applied to the entire face and scalp. Blue light (Blu-U, Sun/DUSA Pharmaceuticals, 10 mW/cm²) exposure to the face and scalp was begun immediately following ALA application, and lasting 30 minutes.

Outcome Measures

The approach involved analyzing photographs and clinical questionnaire responses to test two hypotheses regarding the painless PDT regimen⁸. The first hypothesis was that patients undergoing p-PDT will experience a delayed local inflammatory reaction, characterized by clinically significant erythema and patient-reported symptoms (side effects). Second, we hypothesized that post-PDT inflammatory reactions should be greatest in areas with the highest burden of pre-cancerous cells, namely, in areas with chronic sun damage such as forehead, temples and nose. More specifically, severity of the post-PDT inflammatory response should be positively correlated with AK disease burden, with burden defined as AK lesion counts. The latter hypothesis was fairly straightforward to test because our PDT clinic employs a standardized template wherein AK lesions are recorded in a semiquantitative manner for every patient; see “AK Counts” below.

Erythema Scoring

To be eligible for erythema evaluation, photographs of the face had to be taken under lighting conditions in which periorbital skin was clearly visible and appeared normal. Examples to illustrate the scoring system are shown in Figure 1 and Supplem Figure 1. The periorbital region is always shielded by dark goggles during PDT, and therefore the periorbital area served as a normal control without erythema. Other anatomic sites were judged by a panel of four investigators (A.U, L.H, C.B.W, E.V.M), who scored erythema relative to the periorbital skin. These four evaluators met initially as a group to establish interobserver agreement when scoring the first ten patients. After that, raters evaluated the remaining patients independently. Average erythema scores from all evaluators, for each anatomic site were analyzed (Supplem Table 2). Cohen’s Kappa coefficients were calculated to assess inter-rater reliability (see “Statistical Analysis”).

Figure 1.

An example of erythema evaluation using a semi-quantitative scoring system in a patient with a moderate-to-severe erythema response at 48 hours after p-PDT. For each pre-defined anatomic area (as further described in Supplemental Fig. 1) a value between 0 and 6 was assigned by the physician rater; see Methods. The periorbital skin, which is always protected by dark eye glasses during PDT and therefore remains normal, is defined as zero (0). This photo is from our previous p-PDT trial ⁷ and is shown here for demonstrative purposes; it is not a patient-generated photo used in the study.

Erythema at each anatomic site was scored as follows:

0 – No visible erythema

1 – Pink (barely perceptible) in < 50% of the anatomic area

2 – Pink (barely perceptible) in > 50% of the anatomic area

3 – Mild to moderately red in < 50% of the anatomic area

4 – Mild to moderately red in > 50% of the anatomic area

5 – Bright red in < 50% of the anatomic area

6 – Bright red in > 50% of the anatomic area

NA – Not available, could not be assessed.

For the final analysis, erythema scores were condensed into three levels: mild, moderate, and severe. Mild erythema was defined as scores 0, 1, and 2; moderate erythema as scores 3 and 4; and severe erythema as scores 5 and 6.

Post-PDT Pain Scores

Patients were asked to rate their overall pain at 48 hours after p-PDT on a Visual-Analog Scale (VAS) of 0 – 10 (no pain to extreme pain). For final statistical analysis, the VAS pain scores were condensed into three categories: no pain (0), mild pain (1–3), and moderate to severe pain (4 −10).

AK Counts and Side Effect Evaluation: Use of Composite Scores

For analysis of AK counts, we collected data from our standardized electronic medical record notes, wherein the number of AK lesions is recorded as 0, 1–4 AKs, 5–10 AKs, or >10 AKs. For categorical analysis, these four AK ranges were assigned values of “0”, “1”, “2”, or “3” respectively. Only the values at four sites on the face that most consistently had AK lesions (temples, cheeks, nose, forehead) were used for quantitative analyses; the assigned values for each anatomic site were summed to generate a composite AK score.

For post-treatment side effects, a composite score was calculated by assigning one point for each item in the questionnaire (swelling, peeling, warmth, burning, itching, dryness) reported by the patient. The composite score was defined as the sum of all positive symptoms (0 to 6 possible).

Statistical Analyses:

Statistical analyses were performed using JMP Pro 16.0 (SAS Institute Inc., Cary, NC, USA). Continuous variables were expressed as mean ± SD, and categorical variables as count (%). P-values for continuous variable group comparisons were calculated using student’s t-test. P-values for categorical variable group comparisons were calculated using either Pearson chi-square or Fisher’s exact tests. Fisher’s exact test was used when more than 20% of cells in a contingency table had expected frequencies < 5. Statistical significance was defined by a p-value < 0.05. A Cohen’s Kappa coefficient was calculated to obtain a measure for inter-rater agreement when the average of the four erythema scores, one from each rater, was determined at each anatomic location.

RESULTS

One hundred four patients completed the survey and provided photographs, and 98 patients provided photographs suitable for erythema analysis. Of those, 61 (62.2%) were male, the mean age was 62.2 ± 8.6, the median follow-up time after PDT treatment was 2 days (48 hours), and 69 (70.4%) patients were Fitzpatrick skin type 1, whereas 29 (29.6%) were Fitzpatrick skin type 2. Additionally, 13 (13.3%) patients had a prior history of immunosuppression. Skin cancer history comprised basal cell carcinoma (BCC) in 52 patients (53.1%), squamous cell carcinoma (SCC) in 35 (35.7%), and melanoma in 5 (5.1%). The number and distribution of AKs at the time of p-PDT are displayed in Supplem Table 1 which shows that the majority of lesions were located on the forehead, temples, cheeks, and nose. Erythema scoring of photographs by our four raters (Supplem Table 2) resulted in a mean Kappa coefficient of 0.59 (range: 0.42 – 0.84), corresponding to moderate to near-perfect agreement amongst raters.

Regarding the overall inflammatory response triggered by p-PDT, patient-reported parameters resulting from cutaneous inflammation are shown in Table 1. Overall, most patients experienced 2 or more side effects, including burning, redness, and at least some level of pain. Close to half of the patients (~40–50%) required a topical steroid for relief.

Table 1.

Side effects reported by male and female patients at 48 h post-PDT. ^[a]

	Number of patients (%)
	Male	Female	p-value [b]
Side effects reported
Burning	23 (37.7%)	21 (56.8%)	0.07
Dryness	3 (4.9%)	0 (0.0%)	0.17
Itching	15 (24.6%)	15 (40.5%)	0.10
Peeling	3 (4.9%)	3 (8.1%)	0.52
Redness	59 (96.7%)	37 (100%)	0.27
Swelling	4 (6.6%)	18 (48.6%)	<0.0001
Warmth	6 (9.8%)	9 (24.3%)	0.05
Post-PDT overall pain score (VAS, from 0 to 10)	1.6 ± 1.9	2.9 ± 2.5	0.01 c
Pain Level Reported
None	22 (36.1%)	7 (18.9%)
Mild	29 (47.5%)	16 (43.2%)	0.04
Moderate-Severe	10 (16.4%)	14 (37.8%)
Experienced 3 or more side effects
Yes	34 (55.7%)	32 (86.5%)	0.002
Triamcinolone used d
Yes	24 (39.3%)	18 (48.7%)	0.37

^aThis table illustrates sex based differences in self-reported symptoms. All values are expressed reported as mean ± SD or count (%).

^bP-values were calculated using Pearson Chi-square test, unless otherwise indicated. A value of p < 0.05 is considered significant.

^cStudent’s t-test.

^dPatients were permitted to use triamcinolone 0.1% ointment at any time following the painless-PDT procedure, as needed, to relieve burning or itching.

Interestingly, the data in Table 1 revealed a number of differences between men and women in their post-PDT inflammatory responses. For perceived pain, there was a significant difference between females, who experienced a higher mean pain score (2.9 ± 2.5), compared to males (1.6 ± 1.9; p = 0.01). When pain levels were separated into three pain categories (Table 1), a significant difference was observed (p = 0.04), with females experiencing significantly more moderate-to-severe pain than male patients (37.8% vs 16.4%). In terms of symptoms and signs, females experienced significantly more swelling than males (48.6% vs 6.6%, p < 0.0001), and also more warmth, burning, and itching, although the latter differences were not statistically significant. Looking at side effects more globally, significantly more females experienced 2 or more side effects compared to males (86.5% vs 55.7%, p = 0.002). Finally, when erythema scores were displayed for individual anatomical sites, an overall significant difference in erythema between men and women was observed (Table 2). The data showed that males experienced higher erythema scores than females for the forehead (26.2% vs 2.7%; p = 0.01) and temples (21.3% vs 5.4%; p < 0.0001).

Table 2.

Erythema intensity levels observed 48 h after PDT at individual anatomic sites. ^[a]

		Number of patients (%)
Site	Erythema Intensity	Male	Female	p-value
Forehead	Low	21 (34.4%)	19 (51.3%)	0.01
	Moderate	24 (39.3%)	17 (46.0%)
	High	16 (26.2%)	1 (2.7%)
Temples	Low	17 (27.9%)	27 (73.0%)	<0.0001
	Moderate	29 (47.5%)	7 (18.9%)
	High	13 (21.3%)	2 (5.4%)
	No Value Recorded [b]	2 (3.3%)	1 (2.7%)
Cheeks	Low	11 (18.0%)	4 (10.8%)	0.14
	Moderate	23 (37.7%)	22 (59.5%)
	High	23 (37.7%)	10 (27.0%)
	No Value Recorded	4 (6.6%)	1 (2.7%)
Supralabial	Low	37 (60.7%)	22 (59.5%)	0.05
	Moderate	5 (8.2%)	9 (24.3%)
	High	0 (0.0%)	1 (2.7%)
	No Value Recorded	19 (31.2%) [c]	5 (13.5%)
Nose	Low	12 (19.7%)	9 (24.3%)	0.18
	Moderate	26 (42.6%)	22 (59.5%)
	High	19 (31.2%)	6 (16.2%)
	No Value Recorded	4 (6.6%)	0 (0.0%)
Scalp	Low	4 (15%)[d]	n/a	n/a
	Moderate	16 (59%) [d]	n/a
	High	7 (36%)[d]	1 (n/a)
	N/A (Patients with hair)	34 (n/a)	36 (n/a)

^aThis table identifies sex based differences in erythema scores determined from photographs by a panel of four trained observers (see Methods). P < 0.05 is considered significant.

^b“No value recorded” means that the particular body part was not visible in the photographs.

^cThis value is high because it includes men with moustaches.

^dRelative erythema levels are only reported for bald men.

n/a, not applicable

N.B. Erythema scores from ears were not analyzed, due to extreme variability at that location.

Finally, for assessing whether an association exists between the severity of field cancerization (AK burden) and the post-PDT erythema response, the number of AK lesions versus the intensity of erythema in our 98 patients were compared with a grid distribution approach (Figure 2). For all anatomic sites considered together, the number of AKs showed a highly significant overall association with higher erythema scores, affirming that erythema was positively related to greater AK burden; p < 0.0001. When analyzing the erythema distribution at individual anatomic sites, similar associations between AK burden and erythema response were observed for temples (p = 0.002) and the supralabial area (0.009), and similar positive associations were observed for the nose (p = 0.09), forehead (p = 0.15), and cheeks (p = 0.06); however these did not reach statistical significance.

Figure 2.

Association between AK burden and erythema response. In each AK category (row), the relative distribution of patients between the mild, moderate, and severe erythema category appears to be dependent upon the number of AK. Thus, the erythema intensity distribution shifts to the right as the number of AK lesions increases. Darker color shading indicates a higher frequency of occurrence. P values are from Fishers exact test for comparisons at each individual anatomic site, and from Pearson Chi Square test for the comparison of All Sites.

Additional analyses investigating associations between AK burden and incidence of side effects, and between AK burden and post-PDT pain, failed to reveal significant relationships (data not shown).

DISCUSSION

While skepticism still exists regarding the utility of the p-PDT technique (given the lack of pain induced during treatment), we have demonstrated here that a significant clinical inflammatory response still occurs, albeit delayed with a maximum intensity occurring ~48 hours following p-PDT. Furthermore, the extent of inflammation is positively associated with neoplastic disease burden (i.e., number of AK lesions). These findings provide additional evidence that the p-PDT procedure does in fact induce clinically significant inflammation, which appears to be a key component of the anti-tumor mechanism behind PDT.

Specifically, most patients in our cohort experienced mild or moderate-to-severe pain, along with two or more adverse symptoms at 48 hours following treatment despite being pain-free during the initial p-PDT illumination. Erythema that developed by 48 hours after treatment was most intense in sun exposed areas with chronic sun damage (e.g., forehead, cheeks and nose). The number of AK lesions (which often occur in sun damaged areas, but represents a more specific measure of pre-cancer risk) was shown to be positively associated with erythema intensity. Thus, not only areas of chronic sun damage but also those with a higher degree of skin cancer risk tend to develop a more intense inflammatory reaction.

When stratifying by sex, we found that females experienced more self-reported pain and other symptoms 48 hours following p-PDT than did males. While this may reflect sex-based differences in subjective reporting, or in treatment response due to physiologic differences (with females having thinner skin than males), differences in daily skin care routines and application of topical products might also play a role. Erythema scores also differed between sexes, with males having higher erythema scores in forehead and temple regions. This is likely primarily associated with men having less hair around the forehead and temples; reduced shielding from UV light will result in more chronic sun damage compared to females.

While we did not identify significant associations between AK burden and incidence of side effects or pain post-PDT, greater AK tumor burden was significantly associated with more severe erythema scores at all locations on the face. Thus, the relatively objective nature in which erythema was recorded may serve as a useful indicator of the dose-dependent inflammatory response that occurs with increasing tumor burden following p-PDT.

Multiple limitations exist for this study. Many of the clinical variables collected relied on patient reporting of symptoms following p-PDT, which is inherently subjective. Patient photographs were self-captured (not done in a professional studio), so that the quality of photographs was not completely standardized and therefore the erythema scoring was not totally ideal; however a moderate amount of inter-rater reliability was demonstrated. Furthermore, information in our database was not granular enough to assess long-term clinical outcomes (lesion clearance) in this particular set of patients. However, since the clinical treatment protocol used here was identical to previous studies in which AK lesion clearance was rigorously measured^8,10, the likelihood is high that patients in the current study experienced lesion clearance similar to those in previous clinical trials. In conclusion, despite some limitations, the current data indicate that p-PDT generates a significant and reproducible inflammatory response which targets areas of high AK lesion burden, and almost certainly plays a crucial role in the clinical efficacy of this newly-improved form of field therapy for squamous pre-cancer.

HIGHLIGHTS.

• The mechanism of action of ‘painless’ photodynamic therapy (p-PDT), which can effectively clear facial AK using blue light illumination begun immediately after topical ALA application, is not fully understood.

• Results of this clinical study document the strong inflammatory response that occurs after p-PDT, despite minimal pain during illumination.

• The severity of the post p-PDT inflammatory response is often more intense for women than for men.

• The severity of post-treatment inflammation is correlated with AK lesion burden.

ABBREVIATIONS

AK

Actinic keratosis

ALA

Aminolevulinic acid

PDT

photodynamic therapy

PpIX

protoporphyrin IX

p-PDT

painless photodynamic therapy

ROS

reactive oxygen species

VAS

visual analog score