Regression Analysis of Protoporphyrin IX Measurements Obtained During Dermatological Photodynamic Therapy

Abstract

Photodynamic therapy (PDT) is a light activated drug therapy that can be used to treat a number of dermatological cancers and precancers. Improvement of efficacy is required to widen its application. Clinical protoporphyrin IX (PpIX) fluorescence data were obtained using a pre-validated, non-invasive imaging system during routine methyl aminolevulinate (MAL)-PDT treatment of 172 patients with licensed dermatological indications (37.2% actinic keratosis, 27.3% superficial basal cell carcinoma and 35.5% Bowen's disease). Linear and logistic regressions were employed to model any relationships between variables that may have affected PpIX accumulation and/or PpIX photobleaching during irradiation and thus clinical outcome at three months. Patient age was found to be associated with lower PpIX accumulation/photobleaching, however only a reduction in PpIX photobleaching appeared to consistently adversely affect treatment efficacy. Clinical clearance was reduced in lesions located on the limbs, hands and feet with lower PpIX accumulation and subsequent photobleaching adversely affecting the outcome achieved. If air cooling pain relief was employed during light irradiation, PpIX photobleaching was lower and this resulted in an approximate three-fold reduction in the likelihood of achieving clinical clearance. PpIX photobleaching during the first treatment was concluded to be an excellent predictor of clinical outcome across all lesion types.

Keywords: Ameluz); Metvix); aminolevulinic acid (ALA; dermatology; fluorescence; imaging; methyl aminolevulinate (MAL; non-melanoma skin cancer (NMSC); photobleaching; photodynamic therapy (PDT); protoporphyrin IX (PpIX).

Figure 1

Representative (false color coded) PpIX fluorescence images of an individual superficial basal cell carcinoma (sBCC) (upper back, 70 year old female) (a) before methyl aminolevulinate (MAL) application, (b) three hours later prior to irradiation and (c) immediately following irradiation. Actinic keratosis (AK), sBCC and Bowen’s disease (BD) lesions (d) accumulate more PpIX than is photobleached during MAL-induced PDT irradiation (mean PpIX fluorescence is presented in arbitrary units (a.u.) with bars indicating the standard deviation) and (e) PpIX accumulation and photobleaching (a.u.) is correlated in each licensed lesion type during MAL-PDT, with red dots representing AK, black dots representing sBCC and blue dots representing BD. The lines of best fit are also presented and the R² values for each lesion are: AK 0.47, sBCC 0.45 and BD 0.43; overall R² = 0.45).

Figure 1

Representative (false color coded) PpIX fluorescence images of an individual superficial basal cell carcinoma (sBCC) (upper back, 70 year old female) (a) before methyl aminolevulinate (MAL) application, (b) three hours later prior to irradiation and (c) immediately following irradiation. Actinic keratosis (AK), sBCC and Bowen’s disease (BD) lesions (d) accumulate more PpIX than is photobleached during MAL-induced PDT irradiation (mean PpIX fluorescence is presented in arbitrary units (a.u.) with bars indicating the standard deviation) and (e) PpIX accumulation and photobleaching (a.u.) is correlated in each licensed lesion type during MAL-PDT, with red dots representing AK, black dots representing sBCC and blue dots representing BD. The lines of best fit are also presented and the R² values for each lesion are: AK 0.47, sBCC 0.45 and BD 0.43; overall R² = 0.45).

Figure 1

Representative (false color coded) PpIX fluorescence images of an individual superficial basal cell carcinoma (sBCC) (upper back, 70 year old female) (a) before methyl aminolevulinate (MAL) application, (b) three hours later prior to irradiation and (c) immediately following irradiation. Actinic keratosis (AK), sBCC and Bowen’s disease (BD) lesions (d) accumulate more PpIX than is photobleached during MAL-induced PDT irradiation (mean PpIX fluorescence is presented in arbitrary units (a.u.) with bars indicating the standard deviation) and (e) PpIX accumulation and photobleaching (a.u.) is correlated in each licensed lesion type during MAL-PDT, with red dots representing AK, black dots representing sBCC and blue dots representing BD. The lines of best fit are also presented and the R² values for each lesion are: AK 0.47, sBCC 0.45 and BD 0.43; overall R² = 0.45).

Figure 2

(a) Receiver operating characteristic (ROC) curve shows that PpIX photobleaching predicts clinical outcome at 3 months more successfully than PpIX accumulation. Area under curve (AUC) is 0.88 for PpIX photobleaching and 0.69 for PpIX accumulation and (b) ROC curve for first and second treatment PpIX photobleaching in the patients undergoing two treatments 9 days apart. AUC indicates first treatment PpIX photobleaching is a better predictor of clinical outcome (0.81 versus 0.55).

Figure 3

Scatter plots of age at treatment versus (a) PpIX accumulation (a.u.) and (b) PpIX photobleaching (a.u.) and box plots of (c) PpIX accumulation and (d) PpIX photobleaching distribution for different age categories.

Figure 4

(a) Mean PpIX accumulation (a.u.) and mean PpIX photobleaching (a.u.) recorded during MAL-PDT conducted in different licensed indications (AK, sBCC and BD) on different anatomical locations and (b) comparison of mean PpIX accumulation and mean PpIX photobleaching in acral versus non-acral lesions (AK, sBCC and BD) undergoing MAL-PDT. Bars on both panels represent the standard deviations of the data.

Figure 5

The effect of using air cooling pain relief on the correlation between mean PpIX accumulation (a.u.) and mean PpIX photobleaching (a.u.) during MAL-PDT of AK, sBCC and BD. Black dots represent the pain relief users, with the solid black line representing the line of best fit for pain relief users and the white dots represent no pain relief, with the dashed line representing the line of best fit for no pain relief. The regression coefficients from these models differ (p < 0.05).