)+ %3r )+ %3j]io $ ** &&! $* ) ) :Pl2 - 大阪大学 QZ 99.80# 99.90# 100.00# 100.10# 100.20# 0# 400# 800# 1200# 1600# 2000# 2400# %) Dose#Rate#(MU/min) 10MV#FFFéîâ# ¬~òTN30013#

�)+ � �%3j]io���$�** &"&!��"$* )��) :Pl2��

ÅÕÞąĈTpäî|�}LäîÖà£�H�®¶ô�

º�¦��Û�¬ÝýØ¿î|��L��� �����

! f}�3�[email protected]�

!  �$�** &"&!��"$* )��) ������K.D��! U�[1uSV`a��

�)+ � �%3r��

f}�3�[email protected]�

X�ÏT05z°¡M!#

FFFéîâÏ6MVË10MV#

T0ñ��Õ�n!�#

! f}�3�[email protected]�

!  �$�** &"&!��"$* )��) ������K.D��! U�[1uSV`a��

�$�** &"&!��"$* )��) ������K.D�

�¤jÎ,�#

åîçiJÎ.��

���3rb��

! ØåÑeÀ»�"  � )HH� (�748�O�������%"&�"  )�H� (�748�O�������%"&�"  ¿î��eùþMÍdóĆ�¯I� %'I&%'N�

! t��vL{�w��eÏ©�"  ��dilĀüJ�0,7�3,</08482T��ZcQ�"  Ăý¥¡et��vL�ÿT÷��

! �~�¡eĀüØe¾Ð�"  Ăý¥±Øåe¾Ð�"  �²¢e��¾�

! čëeØå���qs�eê×�

��ò##400MU/min··®�ò2400MU/min#

········20cm×20cmg6£#

··Field/Reference�ÍCC13�nï;300Vð#

vysrb-������

T8/T0�¤jÊ_�·1%��Ê��###

10MV#FFF�

vysrb-������

T8/T0�¤jÊ_�·1%��Ê��###

��ò##400MU/min··®�ò2400MU/min#

········20cm×20cmg6£#

··Field/Reference�ÍCC13�nï;300Vð#

10MV#FFF�

vysrb-�QZ��

99.80#

99.90#

100.00#

100.10#

100.20#

0# 400# 800# 1200# 1600# 2000# 2400#

Rela,ve#Output#(%

)�

Dose#Rate#(MU/min)�

10MV#FFFéîâ#­¬~òTN30013#g6£ò10cmx10cm#

e3bò10cmb#

100MUÕ$�¤j10)e3�

QZ3nvb��¤éãÝÚÛßçÎu�Jò¶2MUÊ2%��#

;5##

;4##

;3##

;2##

;1##

0##

1##

2##

3##

4##

5##

1# 10# 100#

Linearity#(%)�

Monitor#Unit�

4MV(DR=250MU/min)#

6MV(DR=600MU/min)#

10MV(DR=600MU/min)#

6FFF(DR=1400MU/min)#

10FFF(DR=2400MU/min)#

QZ3pcb��¤éãÝÚÛßçÎ�kJò¶2MUÊ.��P0.5%��#

0.0%#

0.5%#

1.0%#

1# 10# 100#

Coefficient#of#varia,on#(%)�

Monitor#Unit�

4MV(DR=250MU/min)#

6MV(DR=600MU/min)#

10MV(DR=600MU/min)#

6FFF(DR=1400MU/min)#

10FFF(DR=2400MU/min)#

����

����

6MV#FFF� 10MV#FFF�

^RFg6£#3cm�40cm#

SSD=100cm¸e3b10cm�

Varian#Installa,on#Product#Acceptance#test#

#FlatnessòSSD100cm¸10cmb¸80%g6£AÊïmax;min)÷2#

#6MV##FFF# #g6£10cm<13% #ï11.0%ð## # # #g6£40cm<40% #ï38.0%ð##10MV#FFF #g6£10cm<20% #ï16.9%ð## # # #g6£40cm<44% #ï42.1%ð�

vyINH<>M-�\��Y���6MV#FFF�

10MV#FFF�0#

10#

20#

30#

40#

50#

60#

70#

80#

90#

100#

;60# ;40# ;20# 0# 20# 40# 60#

Rela,ve#Dose(%

)#

offHaxis#(mm)#

10MV#FB#

10MV#FFF#

3cmx3cm#

4cmx4cm#

5cmx5cm�

3cmx3cm#

4cmx4cm#

5cmx5cm�

0#

10#

20#

30#

40#

50#

60#

70#

80#

90#

100#

;60# ;40# ;20# 0# 20# 40# 60#

Rela,ve#Dose(%

)#

offHaxis#(mm)#

6MV#FB#

6MV#FFF#

���3��FLJ.B�

!  Øå���qs�eðÂ�-  !9<7,64C,>498�70>39/�-  �9=470><4.�D06/�=4C0�-  #08?7-<,�-  �6,>80==M(8�E,>80==N�-  =69:0�-  #0,5�:9=4>498�-  &B770><B�

!  �0,7�;?,64>B�48/0A�!  &?<1,.0�/9=0�!  "?>:?>�1,.>9<�!  �9=0�%,>0�!  �-=96?>0�/9=0�.,64-<,>498�

Definition of parameters for quality assurance of flattening filterfree (FFF) photon beams in radiation therapy

A. Fogliataa)

Oncology Institute of Southern Switzerland, Medical Physics Unit, Bellinzona CH-6500, Switzerland

R. GarciaInstitut Sainte Catherine, Medical Physics Unit, Avignon F-84000, France

T. KnöösRadiation Physics, Skåne University Hospital, Lund S-22185, Sweden and Department of Medical RadiationPhysics, Lund University, Lund S-22185, Sweden

G. Nicolini, A. Clivio, and E. VanettiOncology Institute of Southern Switzerland, Medical Physics Unit, Bellinzona CH-6500, Switzerland

C. KhamphanInstitut Sainte Catherine, Medical Physics Unit, Avignon F-84000, France

L. CozziOncology Institute of Southern Switzerland, Medical Physics Unit, Bellinzona CH-6500, Switzerland

(Received 6 June 2012; revised 11 September 2012; accepted for publication 11 September 2012;published 3 October 2012)

Purpose: Flattening filter free (FFF) beams generated by medical linear accelerators have recentlystarted to be used in radiotherapy clinical practice. Such beams present fundamental differences withrespect to the standard filter flattened (FF) beams, making the generally used dosimetric parametersand definitions not always viable. The present study will propose possible definitions and suggestionsfor some dosimetric parameters for use in quality assurance of FFF beams generated by medicallinacs in radiotherapy.Methods: The main characteristics of the photon beams have been analyzed using specific data gener-ated by a Varian TrueBeam linac having both FFF and FF beams of 6 and 10 MV energy, respectively.Results: Definitions for dose profile parameters are suggested starting from the renormalization ofthe FFF with respect to the corresponding FF beam. From this point the flatness concept has beentranslated into one of “unflatness” and other definitions have been proposed, maintaining a strictparallelism between FFF and FF parameter concepts.Conclusions: Ideas for quality controls used in establishing a quality assurance program when intro-ducing FFF beams into the clinical environment are given here, keeping them similar to those usedfor standard FF beams. By following the suggestions in this report, the authors foresee that the intro-duction of FFF beams into a clinical radiotherapy environment will be as safe and well controlled asstandard beam modalities using the existing guidelines. © 2012 American Association of Physicistsin Medicine. [http://dx.doi.org/10.1118/1.4754799]

Key words: quality assurance, flattening filter free beams, dosimetric parameter definition

I. INTRODUCTION

Conventional medical linear accelerators delivering photonbeams are equipped with a flattening filter (FF) in order toallow delivery of homogeneous dose distributions with broadbeams. This idea was important previously, when conven-tional fields were used for radiotherapy in conjunction with asimple setup, e.g., parallel opposed and four field box tech-niques. The advance of new technologies has led to newmodalities, e.g., intensity modulated therapy (IMRT) with sta-tionary gantry angles, or rotational IMRT (VMAT, volumet-ric modulated arc therapy), or helical IMRT (tomotherapy).These new modalities do not need to produce a flat homo-geneous beam directly. Recently, many studies focused onnonflattened beams, analyzing their characteristics and pos-sible clinical use has been reported.1–3 Flattening filter free

(FFF) beams present unflattened, forward peaked beam, andare available as options on commercial clinically functionallinear accelerators.

Physical and dosimetric differences between standard FFand unflattened FFF beams have been analyzed by vari-ous groups both through measurements and/or using MonteCarlo simulations on modified or clinically available linacs.The main issues are summarized in the paper from Georget al.4 They rely mainly on general beam characteristics,5–14

spectrum, beam energy and depth doses,15–18 backscatter,19

electron contamination,20 out of field dose,10, 11, 20, 21 neutronproduction,20, 22, 23 and shielding requirements.24, 25

Standardized and consolidated beam parameters as de-scribed or referenced in, for example, AAPM TG 142,26 areused today for the quality assurance of flat photon beams, e.g.,flatness, symmetry, and penumbrae. Those parameters are not

6455 Med. Phys. 39 (10), October 2012 © 2012 Am. Assoc. Phys. Med. 64550094-2405/2012/39(10)/6455/10/$30.00

6457 Fogliata et al.: FFF beams quality assurance 6457

FIG. 2. Renormalization point obtained through the profile third derivative.

This is problematic for FFF beams, where the beam profile ofasymmetric jaw setting fields presents strongly different doselevels toward the two beam edges, asymmetrically placed rel-atively to the central axis.

II.A.1. The inflection point

Pönisch et al.7 suggested the use of the inflection point atthe field edge to renormalize a FFF beam to the same doselevel of a FF beam. From this renormalized profile it is thenpossible to evaluate penumbra (as the usual distance between20% and 80% dose levels) and the field size (at the inflectionpoint, or at 50% dose level to keep the common dosimetricfield size definition).

At the inflection point, the second derivative is null (andthe first derivative presents a minimum or a maximum). Theeasiest way to determine the position of the inflection point,not knowing the mathematical expression of the profile, isto plot the dose difference of two adjacent measuring points,!D. The off-axis position of the minimum or maximum at thefield edge represents the inflection point. The location of thispoint is proximal to the 50% for standard beams normalizedto the central axis, and is at the highest gradient, that couldbe of the order of 10%/mm. This means that the position ofthe inflection point can be accurately determined only withvery fine measurement stepping. A common step length usedfor measurements in the penumbra region is one millimeter.With such a precision, together with the detector size andtype in a high gradient region, the dose level that is then usedfor profile normalization (according to Pönisch et al.) couldbe affected easily by a 10% error. This uncertainty value isthen enhanced at the beam central axis, and, for a FFF beamthat has a dose level at the central axis of about 200% withrespect to its corresponding FF beam, the central axis doselevel could vary by up to 40% due to the normalization tothe inflection point. While the measurement stepping can

be finer than 1 mm to reduce possible uncertainty due tothe precision, this is generally not used in common periodicquality assurance measurements nor during commissioning.

II.A.2. The renormalization value

To overcome the uncertainty using the inflection pointmethod, another normalization point should be determined.The renormalization method is conceptually similar to the in-flection point, and comes to the determination of a point in theprofile shoulder of FF beam profiles to renormalize the FFFbeam to the same dose level of the FF beam at that point. The“shoulder point” is located in a shallow dose gradient region,and in a region where the two FFF and FF beams present sim-ilar shapes, before the FFF beams starts to increase in dosetoward the beam central axis. This point could be found as amaximum in the profile third derivative (Fig. 2).

The procedure for its determination is listed here:

– The FFF and the standard FF beam have to be mutuallyaligned in the off-axis direction (both centered relative tothe central axis).

– Normalize the standard beam (FF) as usual, to 100% atbeam central axis (triangle symbol on the right of Fig. 2).

– Compute the third derivative (as !D from measurements)in the penumbra region. This will present two maxima(mimima) in the ascending (respectively descending) pro-file edge. (The third derivative could be obtained fromboth FFF and FF beams, but for consistency the FF beamshould be used).

– The relative dose on the FF profile corresponding to theoff-axis position of the second maximum for the left pro-file edge—closer to the central axis—(first minimum forthe right profile edge) is used to normalize the FFF beamprofile at the same off-axis position (diamond symbol inFig. 2).

Medical Physics, Vol. 39, No. 10, October 2012

�9264,>,�0>�,6��� 0/�#3B=����������

�+*(+*����*')�

�<-,.05�0>�,6�����%"�#���� �����

CC01,#CC13ïIBAð#Blue#Phantom2#(IBA)#

RAMTECsmart#

#

SSD=90cm¸depth=10cm#

GMD=CIzx3vy�����

!  Þt��vLÈgez��[l]h���r~�æãeØåTÀ»[l�-  &548�=:,<482�0G0.>e¾Ð�-  ĉČñÒgeöă èìeûâ�

�

��

��

�

��

��

��

��

��

��

��

� �� �� ��� �� ��� � ��

���

���

��

��%(��""��

�!(�����!(�����!(�����!(������!(�����

��')!('�,�������

�"�*�

����� ��55�

����� ��55�

����� ��55�

������ ��55�

������ ��55�

åìâÖÞæ¯+ÎPDD#

# #ïDmaxÊnormalizeð##

Advanced#Markus#ChamberïPTWð#Tough#Water#Phantom#

RAMTECsmart#

#

SSD=100cm#

Field#size=10cmx10cm#

�

GMD=CIzx3vy�����

!  �406/�&4C0āí²�-  ªĂý¥af��P����-  �Ăý¥af�������

!  &&�āí²�-  ªĂý¥af��P����

��')!('�,���!���'�+����%�#��#�����

���

����

���

���

����

����

�� �� ��� ��� �� �� �� ��

���

����

��

���!��'�+����"��

�!(����

�!(����

�!(����

�!(�����

�!(�����

and 2.5 cm for 10X). No overresponse corrections or biascorrections were conducted for the ionization chamber (20)because only the relative buildup doses for the flat and FFFphotons were of interest in this study. The relative or fractionaldoses were compared between flat and FFF beam to evaluate theirbuildup dosimetric characteristics. At least three repeatedmeasurements were conducted for each beam energy, buildupdepth, and FS to determine the uncertainty values.

Results

Surface dose vs. field size

Good repeatability was demonstrated. All the results were meanvalues, and the standard deviations were less than 2%. The surface

dose within the buildup region was linearly increased with the FS(correlation coefficient R2w1) for both the flat and FFF beams(Fig. 1). The slope was about 1% per cm2 for both 6X and 10X flatand FFF photon beams. Good repeatability of the measurementswas observed. The fractional surface dose, which was surface dose(D0) at any FS divided by Dmax at a 10 ! 10 cm2 field, wasobserved to be greater for the FFF beams than for the corre-sponding flattened beams for both 6X and 10X. This wasconsistent with the larger beam output factors at the phantomsurface for the FFF photons compared with the flattened photonsfor FS " 10 cm. Because the surface dose increased linearly withthe FS, surface dose at was obtained by extrapolation and found tobe 16.4% for 6X flat, 22.8% for 6X FFF, 10.2% for 10X flat, and15.7% for 10X FFF beams.

Doses at different buildup depths

The buildup dose of the 6X FFF photons was modestly larger thanthat of the flattened photons for buildup depths of 0wdmax at FS of2 ! 2 w 10 ! 10 cm2 (Fig. 2). However, the difference was notsubstantial (Table 1). For example, at a buildup depth of 5 mm, thebuildup doses for 6X flat ranged 85% to 87%, which was notsignificantly different from 90% to 91% for 6X FFF photons.

For 6X flattened x-rays with a FS of 10 ! 10 cm2, the frac-tional dose increased from 27% to 94% in the first 7-mm buildupdepth. In comparison, for 6X FFF x-rays, the fractional doseincreased from 33% to 96% in the first 7-mm buildup depth.Therefore, a 6-mm bolus may be adequate for both 6X flattenedand FFF photons to obtain 90% buildup dose on the skin in theapplication of whole-breast irradiation.

Similarly, the buildup dose of the 10X FFF photons was slightlylarger than that of the flattened photons for different buildup depthsat FS of 2! 2w 10! 10 cm2, shown in Fig. 3. Still, the differencewas not substantial (Table 2). For instance, at a depth of 5mm for FSZ 2 ! 2 w 10 ! 10 cm2, the buildup doses for 10X flat and 10XFFF photons ranged from 67% to 72% and 75% to 78%, respec-tively. In the first 2 mm, for 10X flattened x-rays at a FS of 10! 10cm2, the fractional dose increased from 20% to 49%. Similarly, for10X FFF x-rays with a FS of 10 ! 10 cm2, the fractional doseincreased from 24% to 57%.A 1-cm bolusmay be adequate for 10X

Fig. 2. (a) Buildup dose of 6X FFF photons is modestly larger than the flattened photons for different buildup depths 0wdmax at fieldsizes of 2 ! 2 w 10 ! 10 cm2. All data are normalized to dmax z 15 mm at a 10 ! 10 cm2 field. (b) Buildup dose difference between 6XFFF and flattened photons. FS Z field size; FFF Z flattening filter-free.

Fig. 1. Relative surface dose (D0/Dmax) increases linearly withthe field size (w1%/cm2) for both 6X and 10X flat and FFFphoton beams (error bar Z standard deviation). The surfaceoutput factors for field sizes 2 ! 2 w 10 ! 10 cm2 show 6X FFF> 6X Flat > 10X FFF > 10X Flat, which have zero-field-sizesurface doses of 22.8%, 16.4%, 15.7%, and 10.2%, respectively.FFF Z flattening filter-free.

Volume - # Number - # 2012 Surface dose for nonflat x-rays 3

��')!('�,�������%�#��#����

��

��

���

���

��

��

��

�� ��� ���� ���

���

����

��

������"��

�!(�����!(�����!(�����!(������!(�����

*,82���0>�,6�����%"�#������������

wdqe�

!  su�ÆÁ�ڵ˷�-  ½¯su�ÆÁ��-  �úsu�ÆÁ�KFMA_83h��[dāí�

!  ���af�##��9=0�:0<�:6?=0�TÀ»[l]hI��eèìT�UVcl�"  � )�������.�B�:?6=0�O�������� �.�B�:?6=0�,>��7,A�"  )�������.�B�:?6=0�O���������.�B�:?6=0�,>��7,A�

O��b���afßÌdڵ˷eðÂTΰ�

+�*,82�0>�,6������ #�����������

g~WX3���

!  ³�¤jFFFÍÑÓaqS§ïBeam#On#TimeðÎw�#

-  ('y�5}|ͼK��NÎd9#

-  /PÎK�Í5IÆÓ�ÊÎÛìîæÞá&�#

�6,>>080/��%��� �����%�� �

)�'9>,6� (� �������� ����������

(�:0<��406/� ���� �� � ��� ����0,7�98�>470� ��=�F� ��=� � =�F�� =�

�Î*3/¦SBRTï9 ð�

�6,>>080/��%��� �����%� �

� )�'9>,6� (� ���������� �������� ��

(�:0<��406/� ������� �������0,7�98�>470� �=�F�����=� ��=�F����=�

��Î*3/¦SBRT(6 )�

! f}�3�[email protected]�

!  �$�** &"&!��"$* )��) ������K.D��! U�[1uSV`a��

�)+ � �%3���a��!  ClinacÚëîÜ#

-  LinacËLeaf#mo`on#controllerËΧÍ50msΠDS§#

-  MLCβ��>¸Beam;hold...#

!  TrueBeam#

-  SPVÍÑÓ�%�H#µ#5�γ��ï10msð##

#

➡  DMLCͽÂÓ][t�C¿&�#

VRILO279�$"&��03t��!  DMLC;IMRTæêí#-  TrueBeamËClinac#iXÊ_�#

-  �¤jÏ600MU/minÍ�3##

-  Min.#Dose#Dynamic#Leaf#GapÎ�3�Ï��0.20cm#

Clinac##

iX�TB��YòFrac`on#check#(Doselab)�

VRILO279�$"&��03t��!  Fence#Test#

Clinac##

iX�TB�

��,�*)��#"&!���("��)��!  %,:4/,<.Ăý�d ��e¨§d�n\`�,@Tçò�-  ��SjećmØåoÞ¾�-  ¿î�L�oċWlXbTÔQiRd ��Ăý¥SjÜd��¸�eĄïob^`³�Yml�

��,�*)��#"&!���("��)��

vyT�43{|����/;��

!  �r5 #10MV#FFFÕ�nÅÈßÛáæêí#"  �¤�?ò�¾Â�Î0ÀÌ>Ïf»#

"  DVHò�60%�¤+ÊOAR�¤¿�d#

"  MUòÔǾÍ,�ï0.5;1.8MUð#"  JawÏ1.0;3.7cmÎy�A#

Jaw;tracking#(+)�

Jaw;tracking#(;)�

��,�*)��#"&!3^m�[��'!k��

Jaw��C´ó#0.2mm�ßÛáæêí�

���C´ó#0.1MU� Ùíáë)��C´ó#0.2°�

506�

!  �6,>>08482��46>0<��<00�L��-  ò´e�6,>>080/��0,7bõcl�L�Ͳ�-  ¿î��eùþIĂý¥±Øåe¾Ð�-  y�~z���xI$�Ã�eÀ»�-  á¼K���ÍÉe$�Iöă É«²�

�!  '<?0�0,7dil�ø§ÝýØ¿î�-  ���oÊhb[l¤­Çeç»I­Ċ ø§e¹�di^`ik�ø§S_I{�L�~�eÙQ¿îTÄéÓÇ�

���Vs�ÕÆÓͺÈÒÄO7»ÈÉ»È�mRÍ�ÁHxpÅ�ÃÐƹ#

#

��^G·¨¥#

�W��·�m�

±; �·�m�

oQ··´�m�

=<�4·�qL�¢¥�

:F��´´hl-�

�<«1·hl-�

Z�©"·L@�

{�U2·L@�

#

! äëÖíèà×ØìÚÛßçÜ#

�Wv�·\#

<Vª0·\#

XV�#·\#

³oBa·\#

c`Em·\#

�