1. New Theory and Efficacy of Corneal Cross Linking ____________________________
J.T. Lin, Ph.D
New Vision, Inc. Taiwan
3ya-2013

2. Summary Historical of CXL Debating issues (safety, Dresden protocol)
Dynamic of CXL
Kinetic of type-I & type-II
(role of oxygen)
5.Minimum corneal thickness
6. Efficacy & optimal protocols
* References: (papers by JT Lin):
Cornea(8, 2017)
PlosOne (7, 2017)
Ophthalmology Res (7,2017)
IOVS (4,2016); JRS (8,2016)
IJO (10,2015).

3. CXL Milestones (1) Biomechanical studies on KC patient
1980
Biomechanical studies on KC patient
1998
Seiler , University hospital Dresden.and Spoerl starts research on biomechanics of eye tunic.
1994
Improved corneal stability
1995
Evidence of crosslink effect
1996
First experimental result
1997
Therapy test in rabbit
First patient treated
1999
enzymatic resistance after CXL

4. CXL Milestones (2) 2000 First UV LED (370nm) device in market
Study toxic threshold of EC and KC cells
First clinical study published (wollensak et al)
First report about epi-on CXL (Pinelli et al)
2006 Stromal demarcation line (Seiler)
2007 Calculation of diffusion of riboflavin (Spoerl)
2008 CXL by confocal laser scanning microscopy (Mazzotta)
CXL in endothelial decomp, bullous keratopathy (Ehlers)
PACK-CXL (Iseli)
First clinical trial in Australia (Wittig-Silva)
2009 Hypoosmolar riboflavin solution for thin corneas(Hafezi)
Topography guided PRK and CXL (Kanellopoulous)
2010 Investigation of the riboflavin film (Wollensak)
Systemic treatment of keratitis (Makdourni)

5. CXL Milestones (3)
2011
Iontophoresis for riboflavin application (Vinciguerra)
2012
Acce2018lerated CXL AVEDRO (Celik)
2013
Kinetic model of CXL (oxygen dependent) (Richoz),
2014
Antibacterial efficacy of PACK-CXL (Hafezi)
First dynamic model for ribo depletion (Lin)
2015
Laser scan device for focal cross-linking,
First CXL UV-pen device (New vision Inc, patented, Lin)
High power (18, 45 mW) CXL devices (Mlase, Avedro)
2016
First dynamic theory for CXL type-I efficacy profiles
and new criterion for CXL safety (Lin),
FDA approval for Avedro KXL (3 mW). Also CFDA approval.
2017
Kinetic model for type-II CXL (Kling & Hafezi).
Kinetic model for combined Type-I and –II (Lin)
Optimal efficacy and New Protocols (Ln)
First animal study for thin and thick corneas (Kling, Hafezi)

6. Debating issues of CXL:
Safety criterion
[5.4 J/cm^2, 400 um thickness]
(2) Thin corneas (<350 um, safe ?)
(4) Dresden (2003) protocol
(too much RF drops during CXL)
Effective dose (with rib films)
(5) Accelerated CXL
validation of BR law?
(6) Optimal protocol.
(fast and efficient)

7. UV light New dynamic mechanism Experiment setup Power meter
B2 solution (0.005%)
10 mm

8. Measured data * Transmitted light intensity Time (min.)
More UV light transmission for t>0
Dynamic spectra of B2
Transmitted
light intensity
Time (min.)
* JT Lin (published in IJLREC, 2013)

9. Dynamic intensity A(t) = 2.3[ (a-b)C(t) + bCoF +Q]
Our new finding (Lin, pub. IBME, 2013)
I = Io exp (-Az)
A(t) = 2.3[ (a-b)C(t) + bCoF +Q]
diffusion F(z)=1-o.5z/D
conventional
For t=0 only, A=52( 1/cm)
for Co=0.1 %
A=79 to 41, mean 61 (if F=1)
F<1,
z=200 um, A = 67 to , mean 54
z=400 um, A = 55.5 to 37.8, mean 47
Smaller A, higher Intensity.

10. Dynamic of Riboflavin concentration (depletion for t >0)
Initial (t=0)
RF depletion
starts from
surface (z=0).
Time (0 to 120 seconds)
Initial (t

11. CXL Theories developed
Conventional theory
Schumacher, (2012)
Mrochen, Seiler et al
(JOVS, 35:726,2012)
Assumptions:
linear theory, simplified
constant concentration (C)
Errors:
overestimates
effective dose (4.0 J/)
E* = 5.4J/cm^2
Z*= 400 um (cornea thickness)
FIXED parameters
Modern theory
J.T. Lin (2015)
(JRS/2015, JOVS/2016)
Nonlinear,
dynamic theory
of C (z,t)
Safety critera
[C, D, E’, z]
E* = (2.5 – 10.5) J
Z* = (200 – 450) um
C = 0.1 – 0.3%

12. CXL safety formulas (Lin, 2016)
Safety dose
E*=Ed exp(A’z)
Ed: endo cell damage
threshold (at z=400 um)
(2) Minimum Corneal thickness
z* = (1/A) ln (E*/Ed)
(3) Safety Concentration
Co = ln (E*/Ed)/ (2.3mbGz)
For Co = 0.1, 0.2, , 0.5%
(curves 1, 2 ..5)

13. kinetic of CXL type-I & Type-II
Oxygen is required for type-II, but not absolutely needed in type-I .

14. kinetic of CXL type-I & Type-II
Oxygen is required for type-II, but not absolutely needed in type-I .

15. CXL Efficacy type-I & Type-II*
S1 = sqr [ Co/aI] exp (Az)
type-II
S2 = (1-b/[O]) Co [ 1 – exp (-apE)]
Same dose, s.s.
High Intensity, low type-I, Efficacy
Faster Oxygen depletion, but same
tpe-II efficacy as low intensity.
[Q2] profile on surface (z=0) &
at z=100 um. For I = 3,9,10, 30,45 mW

16. Corsslink efficacy (type-I) S = sqr [ (4KCo)/(aI)] exp (Az)
Inverse proportional to the UV light intensity (I)
proportional to RF Co, and depth (z)

17. How to optimize CXL , fast and efficient ?
High power, faster procedure (small T* )
but higher risk
(2) High concentration, deeper CXL depth
(3) Longer illumination time, deeper CXL,
(4) Higher intensity, faster depletion,
but lower efficacy.
How to optimize CXL , fast and efficient ?
OPTIMAL parameters: mW, 20 sec, 0.1%,

18. multi-step-method (MSM) Super-position of Efficacy profiles. Depth (z)
Define: THE ISSUES:
DropF=
Frequency of RF drops (during UV exposure)
Dresden Proto: DropF = 10 to 15 (too much reduction), Effective dose
= 5.4 x0.8= 4. 3 J
(b) Modern proto: DropF = 0,
low efficacy in high intensity
multi-step-method
(MSM)
Super-position of
Efficacy profiles.
Eff = 1 –exp [- (S1+ S2 +S3)]
Depth (z)

19. Proposed Optimal CXL Lin-code 3-24 (8,8,8) 9 - 8 (2.5, 2.5, 3.0)
multi-step-method (MSM)
Hgh effective dose,
Fast procedure,
High efficacy
Lin-code
DropF=2, effective dose =4.3 J (or 5.4x0.8)
(a) I = 3.0 (mW/cm^2)
3-24 (8,8,8)
min.
9 - 8 (2.5, 2.5, 3.0)
(1.3, 1.3, 1.4)
30 – 2.4 (0.8, 0.8, 0.8
Each drop waiting for 1 min. allowing enough diffusion depth.

20. The new trends Combined techniques: sequential or simultaneous
CXL + PRK (2009, same day is better than following day ..)
CXL+ LASIK (2011, 2013, more stable, less aberrations)
CXL+ femtosecond-flap (less regression, prevent corneal ectasia)
CXL+ ICR
CXL + CK (3 months, still have regression)
CXL+ phakic IOL
CXL+ OK (more stable, long term)
CXL (on cornea) to correct low myopia; change of corneal front curvature
(due to thickness deceasing, K value, K=377/r, drops, r increases)
CXL (on sclera) to stop abnormal axial length growth.
CXL+ RGP/OK lens
CXL (on trabecular meshwork, cell pores 3 x increase, glaucoma, Shang-dong U.
Ref: X. Fan. International J ophthalmology 2014,7:

21. CXL-pen (Lin, patented, 2014)
High-power (30-50 mW)
Spot size: 4 to 12 mm
Scanning mode for
customized CXL
KC and PACK

22. IONTOFOR RF device for epi-on
enhance PACK-CXL applying an electrical current
to increase its mobility across a
surface. procedure by reducing the soaking
phase from 30 minutes to 5 min.
No epithelium removal is needed.
電流刺激儀(法國製造)，增強核黃素在角膜內的滲透深度。
同時也能縮短術前等候時間從30分鐘到5分鐘，特別是應用於epi-on的情況 22