Illustrations
Figure 1
Global view and making an exogen-free contrast-enhanced-phototrichogram on human scalp hair in male pattern hair loss.
On the global view (left) we show the region of interest (ROI ; white outline) where we performed a contrast-enhanced phototrichogram with exogen collection (CE-PTG-EC).
Baseline image (t0) shows that all hair is clipped very short in the ROI and that all hair fibres are well contrasted against the background. There is no clipping 48h later while a new image is taken (t2) immediately after extracting the exogen hair.
In the lower panels we show enlarged views with arrowheads of two small fields containing exogen hair. The follicular opening of exogen follicles are pointed with arrowheads: a thin hair in the field (a) with a red contour in t0 is absent at t2 while a thick hair in another field (b) with a blue contour at t0 is removed at t2. The identification of the exogen stage of the hair cycle is documented by the presence of a hair fibre at t0 and its absence at t2. Not shown here is the exogen hair entrapped in a polymerized matrix where its typical terminally differentiated club end can eventually be examined microscopically (for more details see [9] .).
Figure 1
Figure 2
Step by step procedure for scalp hair imaging with CE-PTG-EC, manual processing and growth quantification with computer assisted image analysis.
The first image is taken at t0 (A) while the second is taken 48h later at t2 (B) . The combination after the manually processing, as described in material and methods, in (C) shows the overlay that is made ready for the computer assisted analysis. The details of manual processing at t0 and t2 are shown in the lower panels for t0 (A’ : red lines) and t2 (B’ : arbitrary choice of colours): every single hair fibre is covered by an equally thick calibrated line. The perfect overlay of A’ on B’ is shown on C while the measured difference between t2 and t0, i.e. substraction of A’ from B’, appears in C’ . This results in specific identification of length difference. In c , we adopted a specific colour code for the thickness of fibres: magenta, blue and green were used for thin from ≥20 μm to <40 μm, intermediate from ≥40μm to <60 μm and thick from ≥60 μm to <80 μm, respectively. It is of note that there were no thicker hairs (≥80 μm) in this image. On average, as shown in C’ , the length of thicker fibres exceeeds that of the thinner ones, reflecting faster growth. The measurement of the length divided by the time elapsed between t0 and t2 results in the growth rate for each individual fibre.
Figure 2
Figure 3
Hair length measurement: correlation between experimental data generated by CAIAMP versus Ruler.
There was an excellent and linear correlation between the hair length measurements by CAIAMP (μm (CAIAMP)) i.e. after processing by technicians #B and #C versus rulers (μm (Ruler)). Technician #A was significantly less precise.
Figure 3
Figure 4
Quality control on hair thickness in 3 technicians.
Hair thickness measurements were generated by computer assisted image analysis after manual processing (CAIAMP) by 3 technicians. For each individual hair fibre, differences in thickness (μm) were calculated between CAIAMP and rulers. The measurements of thin (<40 μm) and thick rulers (≥ 40 μm) were analysed separately. Differences in μm were expressed as average % with standard deviation bars, the hair thickness in the rulers being 100% (% μm). Mean difference in thickness (% μm) of fibres processed by technician #A varied significantly for thin as well as for thick hair as compared to the fibre thicknesses contained in the rulers. Technician #C was more precise and accurate in absolute terms (see table 1 ) than #B but the relative differences were not statistically significant between technician #B and #C.
Figure 4
Figure 5
Absolute linear hair growth rate correlates with hair thickness.
Mean hair thickness (x axis ; Thickness (μm)) are distributed in 9 classes starting from 20 μm up to 100 μm. For each thickness class, the mean production or growth rates (y axis; μm/day) are shown for male controls (n = 52 observations in 13 subjects; Control, empty squares) and two groups of men with Male Pattern Hair Loss: 60 evaluations before entry into a clinical trial (Volunteers; n = 20 subjects, plain triangles) and 67 evaluations in non-recruited patients with MPHL (Patients; n = 67, empty diamonds). For detailed statistics see table 2 .
Figure 5
Figure 6
Decreased thickness and linear growth rates are associated with clinical severity in MPHL.
The linear growth rate (y axis, μm/day) is shown in relation to thickness in clinically unaffected healthy males (52 results in Controls) comparatively with Male Pattern Hair Loss patients (n = 67) where the severity was clinically appreciated by using the Hamilton pattern classification. The severity was rated as barely visible (Class I-II; n = 28), clearly receding hairline with moderately affecting the top of the head and vertex (Class III; n = 24) or more severely affecting the top with maintenance of a moderate or very small bridge (Class IV-V; n = 15). In order to maintain a sufficiently large number of observations in each thickness category, we grouped thinner hair (from ≥20 μm to <40 μm), intermediate hair (from ≥40 to <60mm) and thicker hair (≥60 μm).
The more severe the clinical scoring of MPHL (from classes I to V) the more drastic the reduction of LHGR.
Interestingly the preclinical stage, Class I, and the clinically least affected, Class II according to the Hamilton categories, did not show a statistically significant reduction of growth rate whatever the thickness of the hair. Conversely, the growth rate of the thicker hair (≥60 μm) in the most severely affected (Class IV-V) grew at a speed observed in the intermediate hair of the controls (respectively 326 3 versus 323.6 μm/day; detailed statistics are shown in table 3 A ).
Figure 6
Figure 7
LHGR in healthy unaffected Controls and MPHL (Volunteers or Patients) with severity grades III-V.
The growth rates (μm/day) of thinner hair (from ≥20 μm to <40 μm), intermediate hair (from ≥40 to <60mm) and thicker hair (≥60 μm) are shown in controls and two equally severely affected subgroups of MPHL: Volunteers or Patients. Similarly reduced LHGR were present in each of the 3 thickness classes of MPHL subjects as compared with controls.
Figure 7
Figure 8
Complexity of hair dynamics in balding man: the clinical meaning of “hair growth”.
Clinical aspect of the receding hairline two months after clipping (scale bar = 1cm). The variation in hair length might be related to cycling effects for the shorter hair (initial hair regrowth, entry into catagen-telogen) while for the longest hair segments, the situation is more complex. It may reflect the thickness related differences in linear hair growth rate of individual anagen hair follicles, transition from anagen into catagen-telogen, mixed together with other factors such as bending and curliness. Note that a lot of clinically non-significant miniature hair can be seen in the front of the receding hairline.
Figure 8
Tableaux
Auteur
Skinterface and Brussel's Hair Clinic
9 rue du Sondart B 7500 Tournai Belgium
Background: The words “hair growth” frequently encompass many aspects other than just growth. Objectives: Report on a validation method for precise non-invasive measurement of thickness together with linear hair growth rates of individual hair fibres. To verify the possible correlation between thickness and linear growth rate of scalp hair in male pattern hair loss as compared with healthy male controls. Materials and methods: To document the process of validation of hair growth measurement from in vivo image capturing and manual processing, followed by computer assisted image analysis. We analysed 179 paired images obtained with the contrast-enhanced-phototrichogram method with exogen collection (CE-PTG-EC) in 13 healthy male controls and in 87 men with male pattern hair loss (MPHL). Results: There was a global positive correlation between thickness and growth rate (ANOVA; p<0.0001) and a statistically significantly (ANOVA; p<0.0005) slower growth rate in MPHL as compared with equally thick hairs from controls. Finally, the growth rate recorded in the more severe patterns was significantly (ANOVA; P≤0.001) reduced compared with equally thick hair from less severely affected MPHL or controls subjects. Conclusion: Reduced growth rate, together with thinning and shortening of the anagen phase duration in MPHL might contribute together to the global impression of decreased hair volume on the top of the head. Amongst other structural and functional parameters characterizing hair follicle regression, linear hair growth rate warrants further investigation, as it may be relevant in terms of self-perception of hair coverage, quantitative diagnosis and prognostic factor of the therapeutic response.