16. Cervix cancer - The GEC-ESTRO Handbook of Brachytherapy

First page 1 Next page Last page

Animated publication

SECOND EDITION

The GEC ESTRO Handbook of Brachytherapy

PART II: CLINICAL PRACTICE 16 Cervix cancer Li Tee Tan, Kari Tanderup, Jacob Lindegaard, Monica Serban, Remi Nout, Richard Pötter

Editors Bradley Pieters Erik Van Limbergen Richard Pötter

Peter Hoskin Dimos Baltas

Cervix cancer

THE GEC ESTRO HANDBOOK OF BRACHYTHERAPY | Part II Clinical Practice Version 1 - 01/09/2023

16 Cervix cancer

Li Tee Tan, Kari Tanderup, Jacob Lindegaard, Monica Serban, Remi Nout, Richard Pötter

1. Summary 2. Introduction 3. Anatomy 4. Pathology 5. Work up 6. Indications

3 3 4 5 5 6 7

9. Treatment planning

12 17 20 21 25 31 32

10. Dose, dose rate and fractionation

11. Monitoring

12. Results

13. Adverse Events 14. Key messages 15. References

7. Tumour, target volumes and organs at risk

8. Technique

1. SUMMARY

Definitive radiotherapy, comprising external beam radiotherapy (EBRT) with concomitant cisplatin-based chemotherapy followed by brachytherapy (BT), is the treatment of choice for all patients with advanced cervical cancer (FIGO2018 IB3 – IVA). MRI-based image-guided adaptive BT (IGABT) is the new gold standard for cervical cancer BT. The technique allows for individualisation of treatment with dose and volume adaptation, and dose escalation or de-escalation where appropriate, to take into account tumour size and topography at diagnosis and at the time of BT while simultaneously sparing adjacent organs-at-risk (OAR). The entire treatment is predicated on the development of new concepts for target and OAR contouring and dose prescription and reporting, and new adapted applicators with greater flexibility for source placement including combined intracavitary-interstitial capability. A large clinical study of MRI-based IGABT in cervical cancer (EMBRACE-I) has shown unprecedented local and pelvic control rates of 92% and 87%, respectively, across all stages without increasing late morbidity.

2. INTRODUCTION

Standard of care definitive radiotherapy for cervical cancer comprises external beam radiotherapy (EBRT) with concomitant cisplatin-based chemotherapy followed by brachytherapy (BT). For many decades, technological developments in cervix cancer BT have been limited and treatment planning has been based on X-rays and approaches originally developed by the classical BT schools in the early/mid-20th century. The most commonly used system up to now is the Manchester point A system which involves standard doses prescribed to a fixed point regardless of tumour size, topography, and response to EBRT, and doses to organs at risk (OAR). This has resulted in a double penalty of suboptimal local control and survival, particularly for patients with large tumours, and significant treatment-related morbidity, particularly affecting the bowel, bladder, and vagina, with a major impact on health-related quality of life in survivors. With the advent of afterloading equipment in the 1980s, it became possible to modify individual source positions and dwell times to shape the dose distribution to the tumour topography while avoiding adjacent OAR. This has given rise to the development of new applicators with greater flexibility for source placement which would allow better adaptation of dose distributions to create complex shapes. The full potential of these new applicators is critically dependent on the use of 3D imaging, particularly magnetic

Cervical cancer has a low incidence in Western Europe and North America but remains one of the most common malignancies in women worldwide. The human papilloma virus (HPV) plays an important role in the genesis of cervical cancer and is present in >90% of cases. The HPV types most frequently implicated are types 16 and 18 found in ~70% of cases, with types 31 and 45 occurring in a further 10%. While systematic vaccination programmes against the most common high risk HPV types (16 and 18) have been initiated in several countries, they are not available in the vast majority of developing countries where cervical cancer is most prevalent. The survival rate for cervical cancer is 60-70% in Western Europe and North America, compared to only 40-50% in Central/Eastern Europe, and even lower in low- and middle-income (LMIC) countries. The lower survival rate in less affluent countries is largely due to more advanced disease at presentation because of the lack of systematic screening programmes. The availability of HPV self-testing is a potentially game-changing advance for early detection of cervical cancer [1].

Cervix cancer

THE GEC ESTRO HANDBOOK OF BRACHYTHERAPY | Part II Clinical Practice Version 1 - 01/09/2023

resonance imaging (MRI), for improved delineation of tumour and OAR and an understanding of new target concepts of different risk-volumes as proposed by GEC-ESTRO and the International Commission on Radiation Units and Measurements (ICRU). This chapter gives an overview of the concepts and technical developments underpinning the state-of-the-art field of MRI-based image-guided adaptive BT (IGABT) for the treatment of cervix cancer. Strategies to adapt to situations where resources are limited are discussed. Clinical evidence for the improvement in patient outcome resulting from the implementation of IGABT is presented.

The cervix is connected to the bony pelvis laterally by the parametria and is supported by the cardinal ligaments laterally and the uterosacral ligaments posteriorly. Only the posterior part of the cervix is covered by peritoneum (pouch of Douglas). On inspection, the cervix has a central orifice (the external os) with an anterior and a posterior lip. Superiorly, the cervix opens into the endometrial cavity via the internal os at the level of the isthmus. The cervix is divided into the endocervix which is lined with glandular epithelium, and the ectocervix which is lined with squamous epithelium. The size of the cervix varies with parity and age but is approximately 2-3 cm in both diameter and length. The total length of the uterine cavity from the external cervical os to the fundus typically varies from 4-12 cm. The regional lymph nodes (LN) for the cervix are parametrial, obturator, internal iliac, external iliac, common iliac and pre sacral. Para-aortic LN involvement was previously classified as distant metastasis in FIGO 2009 and TNM staging but has been reclassified as Stage IIIC2 disease in FIGO 2018 (Table 1).

3. ANATOMY

The uterus lies between the bladder anteriorly and the rectum posteriorly and is divided into the uterine corpus superiorly and the cervix inferiorly, connected by a narrow portion known as the isthmus. The position of the uterus is typically anteverted and anteflexed, but it may also be straight or retroflexed.

TABLE 1 COMPARISON OF 2009 AND 2018 FIGO STAGING CLASSIFICATION. Stage 2009 FIGO definition

2018 FIGO definition

Confined to cervix

≤5 mm depth and ≤7 mm width

≤5 mm depth ≤3 mm depth

IA1 IA2

≤3 mm depth

>3 mm and not >5 mm depth

>3 mm and ≤5 mm depth

>5 mm depth

IB1 IB2 IB3

≤4 cm greatest dimension >4 cm greatest dimension

≤2 cm greatest dimension

>2 cm and ≤4 cm greatest dimension

>4 cm greatest dimension

Beyond uterus, but not involving lower third vagina or pelvic sidewall

IIA

Upper two-thirds vagina

IIA1 IIA2

Upper two-thirds vagina and ≤4 cm Upper two-thirds vagina and >4 cm

IIB

Parametrial invasion

III

Lower third vagina, pelvic sidewall and ureters

Lower third vagina, pelvic sidewall, ureters and lymph nodes

IIIA IIIB IIIC

Lower third vagina

Pelvic sidewall or ureters

Pelvic and para-aortic lymph node involvement

IIIC1 IIIC2

Pelvic node involvement

Para-aortic node involvement

Adjacent and distant organs Rectal or bladder involvement Distant organs outside pelvis

IVA IVB

Cervix cancer

THE GEC ESTRO HANDBOOK OF BRACHYTHERAPY | Part II Clinical Practice Version 1 - 01/09/2023

4. PATHOLOGY

including anaemia and renal dysfunction also have a negative impact on prognosis. Gynecological examination is an essential part of tumour assessment and is ideally carried out jointly by the gynecological surgeon and the radiation oncologist, preferably under general anesthesia. Pelvic examination begins with inspection of the external genitalia and insertion of a speculum to visualise the ectocervix. Next, the extent of disease on the cervix and vagina are digitally assessed per vagina; bimanual abdomino-vaginal examination is useful to assess the size and mobility of the uterus. Pelvic examination is completed by bi-digital rectovaginal examination to assess for parametrial involvement (right hand right pelvis, left hand left pelvis). Cystoscopy and/or proctoscopy are indicated if there is suspicion of organ infiltration on imaging. In all cases, a biopsy should be obtained for histology. Cross sectional imaging often provides additional information on local tumour extent that is not available from clinical examination. MRI is the imaging modality of choice because of its superior soft tissue delineation. MRI is superior to clinical examination for detecting early proximal parametrial and uterine involvement (Figure 2a). However, distal parametrial and vaginal involvement

The most common histological type is squamous cell carcinoma accounting for 80-85% of cases, followed by adenocarcinoma which is more common in young women. Non-squamous tumours are generally considered to have a worse prognosis but at present, are treated in the same way as squamous cancers. Different macroscopic forms are described (Figure 1) which may have an impact on prognosis.

5. WORK UP

Optimal management is critically dependent on thorough assessment of patient and tumour factors. Patient comorbidity and performance status will influence tolerance of treatment and risk of side-effects. Smoking has been shown to adversely affect response to treatment and increase risk of acute and late toxicity. Laboratory findings

Figure 1. Examples of macroscopic types a. Sagittal MRI showing an exophytic tumour arising from the cervix – at examination under anaesthetic, there was no evidence of vaginal involvement. b. Axial MRI showing an expansive Stage IIIB tumour. Note that the lateral edge of the parametrium is ~15 mm from the bony pelvis c. Axial MRI showing an infiltrative Stage IVA tumour invading into bladder (confirmed on cystoscopy and biopsy)

Figure 2. Primary tumour on MRI a. Stage IIB disease. There is breach of the dark stromal ring on the right indicating proximal parametrial involvement. b. Stage IIIB disease. Extensive left parametrial involvement with hydronephrosis. Distal parametrial involvement may be less obvious on MRI compared to clinical examination.

Cervix cancer

THE GEC ESTRO HANDBOOK OF BRACHYTHERAPY | Part II Clinical Practice Version 1 - 01/09/2023

are better assessed on clinical examination (Figure 2b). Transvaginal or transrectal ultrasound (US) has been shown to be as sensitive as MRI for detecting parametrial disease in expert hands. Computed tomography (CT) is less useful than clinical examination for assessing local tumour extent but is equal to MRI for assessing LN involvement. Positron emission tomography (PET) in combination with MRI or CT has been shown to have greater sensitivity and specificity for detecting LN metastases than MRI or CT alone. Comparison of the different imaging modalities is shown in figures 3 and 4. The clinical findings and imaging findings should be accurately and systematically recorded by the examining physician. Examples of diagrams developed for this purpose are shown in figure 5. The FIGO staging system is the first staging system developed to compare treatment results of cervical cancer. Traditionally, it was mainly based on clinical findings by gynaecological examination and did not consider LN status. In 2018, the FIGO system was revised [2, 3] to include any LN involvement regardless of local tumour extent as substages of Stage III disease (Table 1); however, in doing so, some information about local tumour extent has been obscured. The TNM classification defines the status of the primary tumour and LN separately unlike the FIGO staging system but up to now, has not been extensively used making it less useful for comparisons with historical results. The T score is a new proposal for assessing the primary tumour burden involving a simple scoring system which integrates the findings from clinical examination

and MRI [4]. Evidence from the EMBRACE-I study [5] has shown that this system provides additional prognostic information on local control and survival based on the pattern of local spread at diagnosis and the degree of regression observed at the time of BT.

6. INDICATIONS

Guidelines for the multidisciplinary management of cervical cancer were issued by European Society of Gynaecological Oncology (ESGO), ESTRO and European Society of Pathology (ESP) in 2018 [6]. Definitive radiotherapy is recommended as the treatment of choice for all patients with advanced cervical cancer defined as local disease FIGO2018 IB3–IVA and/or LN involvement. Radiotherapy is also an effective alternative to surgery for patients with limited disease (IB, ≤4 cm). The combination of surgery with post-operative EBRT should be avoided as it has been shown to increase morbidity without improving survival. Brachytherapy alone is a curative option in patients with limited disease who are not suitable for surgery or EBRT. Preoperative BT followed by surgery is used in a limited number of centres and should only be used by teams experienced in this approach. The standard protocol for definitive radiotherapy is EBRT (preferably with intensity-modulated radiotherapy (IMRT)), with concomitant cisplatin chemotherapy followed by BT. Brachytherapy is crucial

Figure 3. Comparison of MRI, CT and PET-CT for primary tumour

Figure 4. Comparison of MRI, CT and PET-CT for involved node

Cervix cancer

THE GEC ESTRO HANDBOOK OF BRACHYTHERAPY | Part II Clinical Practice Version 1 - 01/09/2023

Figure 5. (A) Example of diagram for recording findings of clinical examination (at diagnosis and at brachytherapy) (From www.embracestudy.dk) On the direct en face view, the shaded area represents the vaginal fornix.

a. Line separating upper third from middle third of vagina b. Line separating middle third from lower third of vagina

(B) Revised diagram for recording findings of clinical examination for CT-based contouring (From [13]) These clinical diagrams were revised to include measurement scales in all directions to facilitate recording of tumour target dimensions in a reliable and reproducible way. As parametrial disease is difficult to visualise on CT, a new parameter - “Near Maximum Distance” (NMD) - has been introduced for more precise width assessment on clinical examination.

for optimal local control and cure and attempts to replace BT with high tech EBRT techniques (e.g. IMRT or stereotactic radiotherapy) have resulted in inferior outcomes [7]. The use of image guidance has facilitated the use of BT in all stages of disease, including Stage IVA even with bladder and/or rectal fistula. Similarly, BT has an increasingly important role to maximise local control of selected patients with metastatic disease (good performance status, low volume metastases, good response to systemic chemotherapy).

7. TUMOUR, TARGET VOLUMES AND ORGANS AT RISK

In 2000, the GEC-ESTRO GYN Working Group was established to support and shape the emerging field of gynaecological IGABT. Clinicians from a few pioneering European IGABT centres (Leuven, Paris, Vienna) with different historical traditions met to discuss and agree a common language for prescribing, recording and reporting IGABT for cervix cancer. This culminated in the publication of two recommendations on contouring and dose/volume reporting in 2005 and 2006 [8, 9]. In 2005, the GEC-ESTRO GYN Working Group founded a network to promote collaboration between the increasing number of institutions with research and development activities in IGABT. The group launched the “IntErnational study on MRI-based BRachytherapy in locally Advanced CErvical cancer” (EMBRACE-I, www.embracestudy.dk) to evaluate the outcome of IGABT in a multi-centre setting in 2008. In 2010 and 2012, the GEC-ESTRO GYN network published a further two recommendations on applicator reconstruction and imaging [10, 11]. The Gyn GEC-ESTRO Recommendations I-IV have been used as the conceptual framework for the implementation of IGABT worldwide and are embedded into the new ICRU Report 89

Figure 6. General tumour and adaptive target concepts – schematic diagram General adaptive target concept with a CTV-T init based on the GTV-T init , and an individualised CTV-T adapt (CTV-T HR) based on the GTV-T init response after treatment and assessment of GTV T res . (From ICRU 89)

“Prescribing, Recording, and Reporting Brachytherapy for Cancer of the Cervix” [12]. In 2021, the GEC-ESTRO/ICRU concepts were adapted for CT-based contouring [13].

7.1 Tumour and target volumes The ICRU89/GEC-ESTRO recommendations are based on repetitive tumour assessment through clinical examination and cross-sectional imaging, with adaptation of dose according to the tumour extent at diagnosis, and the response to EBRT at the time of BT (Figure 6). A common terminology for different target volumes with different risks of recurrence at different time points has been defined.

Cervix cancer

THE GEC ESTRO HANDBOOK OF BRACHYTHERAPY | Part II Clinical Practice Version 1 - 01/09/2023

Figure 8. Adaptive target concepts for a stage IIB bulky cervical cancer at brachytherapy (same patient as in Fig 7) Adaptive target volumes at time of brachytherapy: CTV-T_HR based on the tumour burden after chemo-radiation i.e. the GTV-Tres and any residual pathologic tissue (stripes), and a CTV-T_IR based on the tumour burden at diagnosis i.e. the GTV-Tinit. The CTV-T_LRinit represents the compartments of potential tumor spread at diagnosis (Modified from ICRU 89)

Figure 7. Initial target concepts (EBRT/IMRT) for a stage IIB bulky cervical cancer at diagnosis Initial GTV-T and CTV-T: large GTV-T init ,, initial CTV-T HR, and initial CTV-T LR: coronal, transversal and sagittal view. (Modified from ICRU 89 and EMBRACE II protocol, www. embracestudy.dk).

Figure 9. GTV-T res at time of brachytherapy The GTV-T res is defined as residual high signal abnormality on T2-weighted MRI. Palpable induration in the parametria should not be included as this cannot be easily distinguished from radiation fibrosis.

At time of diagnosis, the following volumes are defined (Figure 7): • Initial gross tumour volume (GTV-Tinit)- the GTV-Tinit is the primary tumour at diagnosis as assessed by clinical examination and imaging. The clinical and imaging volumes may be different, and a composite GTV-T should be delineated in these situations. • Initial high-risk clinical target volume (CTV-T_HRinit) - the CTV-T_HRinit is the volume bearing the highest risk of recurrence at diagnosis. As a minimum, this includes the whole cervix in addition to the GTV-Tinit. • Initial low-risk clinical target volume (CTV-T_LRinit) – the CTV-T_LRinit represents compartments at risk for potential microscopic spread from the primary tumour. In locally advanced cervical cancer, the CTV-T_LRinit comprises the whole

parametria, the whole uterus, the upper part of the vagina, and the anterior/posterior spaces towards the bladder and rectum. For definitive radiotherapy, additional adaptive volumes are defined at the end of initial chemo-radiation with 45-50 Gy (which is assumed sufficient to control microscopic disease), i.e. at the time of BT. The relevant volumes (Figure 8) are: • Residual gross tumour volume (GTV-T res ) - residual tumour at the time of BT. • Adaptive high-risk clinical target volume (CTV-T_HR) – volume bearing the highest risk for recurrence. • Intermediate-risk clinical target volume (CTV-T_IR) – accounts for original GTV-Tinit at diagnosis

Cervix cancer

THE GEC ESTRO HANDBOOK OF BRACHYTHERAPY | Part II Clinical Practice Version 1 - 01/09/2023

TABLE 2 STRATEGIES FOR CT-BASED CONTOURING Table showing four imaging strategies for CT-based contouring. Strategy 1 has the greatest contouring uncertainty while strategy 4 has the least. For each strategy, the addition of real-time transrectal ultrasound during brachytherapy may reduce contouring uncertainty. Careful clinical examination with precise documentation of findings is key for all strategies (see figure 5b) (Courtesy of Umesh Mahantshetty) At diagnosis Before brachytherapy

With applicator

1 2 3 4

CT CT CT CT

MRI

MRI MRI

MRI

The development and evolution of these concepts are described in detail in chapter 5 of ICRU report 89 [12]. Originally, these target concepts were developed for contouring on MRI. However, the limited availability of MRI in LMIC countries have given rise to adaptation of the concepts for contouring on CT [13] and US. A summary of key points regarding the contouring of these volumes using different imaging modalities is given below. Residual gross tumour volume (GTV-T res ) The GTV-T res is defined as residual high signal abnormality on T2 weighted MRI (Figure 9). Unlike the GTV-Tinit, the GTV-T res does not include palpable induration in the parametria as this cannot be easily distinguished from radiation effect. As a result, the GTV-T res cannot be reliably contoured if CT alone is used apart from visible disease on the cervix and/or vagina. Adaptive high-risk clinical target volume (CTV-T_HR) The CTV-T_HR consists of the GTV-T res , the whole cervix even if uninvolved, and any extra-cervical residual pathologic tissue as defined by clinical examination and on MRI. The residual (extra cervical) pathologic tissue is defined as one or more of the following: residual palpable mass, residual visible mucosal change, pathologic induration, residual grey zones, any other residual pathologic tissue on MRI or clinical examination. To be included as residual pathology, these regions must be located inside the GTV-Tinit and have characteristics that are consistent with residual disease and direct access to the diagnostic MRI during BT contouring is highly recommended. Such pathologic tissues may include parts of the cervix, parametria, uterine corpus, vagina, rectum, or bladder, according to the initial spread of disease. All suspect areas of residual disease should be included in delineation of the CTV-T_HR, with no margins added. CT-based contouring of the CTV-T_HR tends to result in overestimation of the width of the target [14] while the superior extent is not discernible. Various strategies have been suggested to minimise contouring uncertainties [13] based on the clinical remission pattern after EBRT and information available from supplementary imaging modalities (Table 2). Careful clinical examination with precise documentation of findings is key regardless of strategy – a revised diagram for recording clinical findings is provided in the IBS-GEC ESTRO-ABS recommendations for CT based contouring in IGABT for cervical cancer [13] (Figure 5b). Intermediate-risk clinical target volume (CTV-T_IR) The GTV-Tinit carries the highest density of tumour cells at the start of treatment. It can be assumed that residual tumour cells

Figure 10. Tumour and target concepts (MRI-based IGABT) MRI at diagnosis (left panels) and at time of brachytherapy with the applicator in situ (right panels). The brachytherapy targets (blue: GTVres, red: CTV_HR, green: CTV_IR) should be contoured on para-axial slices (upper right panel) and inspected for consistency in the sagittal sequence (lower right panel). The MRI at diagnosis (left panels) is used to identify grey zones and to ensure that the CTV_IR contour fully covers the primary tumour extension. Note that the CTV_IR has been edited to exclude areas where microscopic disease cannot extend directly (e.g. in peritoneal cavity outside the uterus). (From EMBRACE II protocol, www.embracestudy.dk)

will remain at various locations in this volume, even if the tissue has changed to a macroscopic normal appearance on clinical examination and imaging. The concept of the CTV-T_IR was developed to take the GTV-Tinit into account as superimposed on the topography at the time of BT. Special attention is particularly needed for cases with vaginal extension at diagnosis to ensure full coverage by the CTV-T_IR. In addition, the CTV-T_IR takes into account that there is likely to be significant microscopic tumour burden outside the CTV-T_HR by adding a margin surrounding the CTV-T_HR like a shell even where there was no initial GTV-T. The GEC ESTRO recommendations suggest a 10 mm margin in the lateral and cranio-caudal directions and 5 mm in the anterior–posterior direction. However, the CTV-T_IR should not include areas where microscopic disease

Cervix cancer

THE GEC ESTRO HANDBOOK OF BRACHYTHERAPY | Part II Clinical Practice Version 1 - 01/09/2023

Figure 11. Variation in utero-vaginal location of applicator and its impact on location of high dose areas in adjacent OARs Schematic anatomical diagrams (sagittal view) showing variation in location of OAR volumes of interest (2 cm 3 and 0.1 cm 3 ) depending on patient anatomy, applicator position and OAR topography. (From ICRU 89)

cannot extend directly e.g., in the lumen of involved bladder and rectum or in the peritoneal cavity (Figure 10). For CT contouring, the CTV-T_IR can be generated with safety margins based on the CTV-T_HR in line with the GEC ESTRO/ ICRU 89 recommendations. For optimal tumour and target contouring, the contouring physician should have performed the gynaecological examination prior to insertion of the applicator. Information from the clinical drawings from gynaecological examination at diagnosis and at the time of BT, as well as MRI at diagnosis and at time of BT with the applicator in situ, should be available at the contouring station. Contouring of tumour volumes should be performed for each insertion of BT applicators on T2-weighted para-axial MRI sequences; para-coronal and sagittal sequences should be inspected during the process to ensure consistency of target contouring in these sequences. The MRI-based target delineation can be superimposed for subsequent fractions of BT if only CT with the applicator in place is used. 7.2 Organ at risk and morbidity concepts Several morbidity endpoints and OAR of interest have been identified based on the typical morbidity profiles seen with cervical cancer radiotherapy. Historically, publications have focussed on the bladder, rectum and bowel. In recent years, there has been increasing interest in various symptoms for a given organ (e.g., rectum: bleeding, proctitis fistula) and in other OARs such as the vagina (sexual outcome), sigmoid and ureter and also in general symptoms (e.g., menopausal symptoms, pain, fatigue). There is also recognition that different specific morbidity endpoints are related to different sub-volumes or points within each OAR (e.g.,

Figure 12. OAR contouring and GTV res , CTV_HR, CTV_IR The outer contour of each OAR should be contoured - bladder (yellow), rectum (brown), sigmoid (orange), bowel (light green). (From EMBRACE II protocol, www.embracestudy.dk)

Cervix cancer

THE GEC ESTRO HANDBOOK OF BRACHYTHERAPY | Part II Clinical Practice Version 1 - 01/09/2023

Figure 13. Examples of applicators (Reprinted from Semin. Radiat. Oncol., 29/3, Tan LT, Tanderup K, Kirisits C, et al, Image-guided Adaptive Radiotherapy in Cervical Cancer, 284-298, 2019, with permission from Elsevier).

8. TECHNIQUE

for bladder: bleeding and cystitis and urinary frequency and urinary incontinence). Typical BT-related morbidity, such as telangiectasia, ulceration and fistulae, tend to be related to the dose received in small absolute volumes (e.g., 2 cm 3 , 0.1 cm 3 ). These volumes often have different locations in individual patients depending on vagina and OAR topography and applicator location (Figure 11). For small absolute volumes of 2-3 cm 3 , delineation of the outer contour of the OAR has been shown to be sufficient (Figure 12). Applicator-related and anatomy-based reference points (see 10.1.1) have also been shown to be highly predictive for some OAR morbidity. Consistent protocols for filling status of a hollow OAR is essential for comparison between centres. For bladder filling, various clinical protocols have been suggested, none of which have proven superiority. One suggested protocol is to keep the transurethral catheter on open drainage to provide an “empty” bladder [15]; this is the only practical strategy for PDR. For HDR, another option is to instil a defined volume (e.g. 50 cm 3 ) after emptying the bladder [16]; this should be carried out before the planning MRI/CT and immediately prior to each treatment. For the recto-sigmoid and other bowel, the filling status is similarly important although practice in terms of protocols is even more variable. A suggestion is to use a pre-intervention enema and a rectal tube to achieve an empty recto-sigmoid colon; this is particularly relevant for PDR treatments because of the extended treatment period. Organ motion between imaging and dose delivery may lead to discrepancies between prescribed and delivered absorbed dose. The assumption of static anatomical location of hotspots is recommended for small volumes in fractionated BT to assess the worst-case scenario accumulated high-dose region for a particular treatment.

8.1 Applicator selection Current applicators for cervix BT all have an intrauterine (IU) component (the tandem) and an intravaginal component. The most common intravaginal components are based on classical techniques i.e. ring (Stockholm) or ovoids/colpostats (Manchester, Fletcher). In subsequent developments, a single plane of straight interstitial (IS) needles was incorporated into the ring or ovoid to encompass lateral extension of disease into the parametrium more fully. Latest developments in applicators have incorporated a double plane of IS needles (straight and oblique). Some applicators have also been adapted to allow treatment of disease extension down the vagina, the so-called universal applicator. A few centres create individualised applicators for some patients to accommodate difficult tumour topography – examples include the mould technique [17] and 3D printed templates for transvaginal needle insertion [18]. Examples of some types of applicators are shown in figure 13 [19]. The selection of applicator is based on clinical examination and 3D imaging performed prior to or during the BT application taking into account the definition of the clinical target volume(s). In most centres, the type of intracavitary (IC) applicator (e.g., tandem and ring, tandem and ovoid, or tandem with a vaginal mould) is determined by the preference of the radiation oncologist and availability of equipment. However, if several applicator types are available, the type of vaginal applicator may be selected based on the vaginal geometry. The appropriate size of ring or ovoids is based on clinical examination of the vagina – the largest ovoids, ring, or cylinder that fit comfortably into the vaginal apex abutting the cervix should be chosen. Inserting a vaginal applicator that is too large

Cervix cancer

THE GEC ESTRO HANDBOOK OF BRACHYTHERAPY | Part II Clinical Practice Version 1 - 01/09/2023

Prior to applicator insertion, a urinary catheter is inserted into the bladder for patient comfort throughout the duration of the implant and to allow calculation and reporting of the dose to the bladder neck according to the ICRU definition of the bladder point [12, 26]. The catheter balloon is inflated with radiopaque solution (7 cm 3 ) and should be pulled down towards the base of the bladder until it is fixed at the bladder neck and fixed outside. After applicator insertion, the vagina is packed tightly with gauze to keep the applicators in position and to displace the bladder and rectum; in some applicators, a separate rectal retractor can also be attached. For larger residual disease and cases of unfavourable anatomy, combined IC/IS applicators with additional lateral (straight +/- oblique) needles are advantageous. These applicators have holes in the ovoids or rings to guide needle placement. The use of blunt needles is preferred to minimise complications from needle insertion. The planned depth of needle insertion in the parametrium is determined by the cranial extent of CTV-T_HR, usually 4–5 cm above the upper surface of the vaginal applicator. Pre-BT imaging can be used to determine the optimal length and location of the needles to be inserted while intra-operative transrectal US can be used to guide real-time needle placement. The goal of treatment planning is to obtain the best possible chance for an uncomplicated cure of the individual patient. Treatment planning is a multi-step process predicated on pre-defined planning aims and dose constraints for the combined dose distributions of EBRT and BT based on information available at diagnosis. BT in particular offers many opportunities to achieve the defined dose planning aims. Variations of the implant geometry, loading pattern and time–dose pattern (dose rate, fractionation) allow modification of the therapeutic window to deliver an optimal absorbed dose for the targets while sparing the OAR. With the introduction of BT afterloading machines, the historical radium systems were converted into a pattern of dwell positions and times for a single stepping source. The so-called “standard loading patterns” of modern BT were introduced to reproduce isodose distributions of the historical Manchester/Fletcher or Stockholm BT schools. 9.1 Combining EBRT and BT: dose and fractionation strategies The 3D dose distribution in definitive radiotherapy for cervical cancer is characterised by exposure of relatively large tissue volumes to intermediate dose levels from EBRT, and exposure of smaller volumes to high doses in the region of the BT boost and EBRT boosts to involved lymph nodes. The EBRT and BT doses are converted into equi-effective dose according to the linear quadratic (LQ) model to facilitate dose accumulation across the two modalities. The current standard for reporting equi-effective dose in cervix BT is equivalent dose in 2Gy fractions (EQD2) using α/β ratios of 10 Gy for tumour and 3 Gy for OAR. For PDR BT, a repair half time of 1.5 h is used. It is important to note that the characteristics of the combined EBRT and BT dose distribution are highly dependent on the weighting between EBRT and BT. The radiobiological aspects of BT are discussed in more detail in chapter 7 of ICRU Report 89 for cervix cancer and in the radiobiology chapter of this handbook (Chapter 5). 9. TREATMENT PLANNING

may result in the entire implant lying lower in the vagina which in turn may make it impossible to load the vaginal applicator. The addition of IS needles can compensate for a smaller sized vaginal applicator in patients with unfavourable anatomy. Serban et al. [20] compared the dosimetry of tandem-ring vs. tandem-ovoid treatments within the EMBRACE-I study. For similar point A doses with IC implants, the mean CTV-T_HR D90% was 3.3 Gy higher and V85Gy was 23% lower for ring compared with ovoid centres (at median target volumes of 30 cm 3 ). The mean bladder/rectum doses (D2 cm 3 and ICRU-point) were also 3.2-7.7 Gy lower with the ring although the dose to the vaginal 5-mm lateral point was 19.6 Gy higher. The addition of IS needles resulted in increased CTV-T_HR D90% in large target volumes of 60 cm 3 for both ring and ovoid centres (mean 8.9 Gy and 5.4 Gy respectively) while V85Gy and OAR doses were generally unchanged (ring) or reduced (ovoid). IC/IS implants therefore improve target dose as well as dose conformity compared to IC implants. 8.2 Applicator insertion Modern BT applicators are made of MR-compatible plastic or CT-compatible titanium or tungsten. The IS needles are mostly disposable plastic. Increasingly, the manufacturers have tended to reduce the number of options available in terms of length and angle of the IU tandem, and size and form of the vaginal component to minimise costs. Insertion of applicators is usually performed under spinal or general anesthesia. Less common strategies include oral or intravenous sedation with a local paracervical block [21-23] and hypnosedation [24]. The patient is positioned in the lithotomy position. At each insertion, the radiation oncologist begins with a thorough gynaecologic examination to assess the patient anatomy and extent of tumour. Vaginal specula are introduced to visualise the cervix, and one or two cervical forceps are placed on the anterior lip of the cervix. The length of the uterine cavity is measured using a semi-flexible or metal probe. The cervical os may be difficult to locate if there is extensive tumour destruction. Uterine perforation and false passages are recognised complications of applicator insertion, especially if the cervical os is difficult to identify due to tumour destruction or if the uterus is not in the standard anteflexed anteverted position. Transabdominal or transrectal US are useful to guide correct placement of the tandem within the uterine canal; a recent randomised controlled trial [25] showed that use of transabdominal US reduced the incidence of perforation from 12.5% to 1.25%. There are no standard recommendations for the management of uterine perforation - the clinician has to decide whether to proceed or abandon treatment; intravenous or oral antibiotics are usually prescribed. After the uterine canal is measured, the cervical canal is dilated to accommodate the tandem. With modern applicators, the diameter of the tandem is often small (~3 mm) and significant dilatation is not required. An IU tandem of the appropriate length and curvature is inserted through the cervical os into the uterine cavity. The intravaginal component of the applicator is chosen depending on tumour diameter and topography and patient anatomy. If the anatomy is very narrow, an IU applicator extending into the vagina may be used alone or in combination with a vaginal cylinder.

Cervix cancer

THE GEC ESTRO HANDBOOK OF BRACHYTHERAPY | Part II Clinical Practice Version 1 - 01/09/2023

TABLE 3 RISK-ADAPTED BRACHYTHERAPY DOSE PRESCRIPTION PROTOCOL FOR EMBRACE-II Target D 90 CTV HR EQD2 10 D 98 CTV HR EQD2 10 D 98 GTV res EQD2 10

D 98 CTVI R EQD2 10

Point A

>90 Gy <95 Gy

Planning aims

>75 Gy

> 95 Gy

> 60 Gy

>65 Gy

Limits for prescribed dose

>85 Gy

> 90 Gy

ICRU recto- vaginal point EQD2 3

Bladder D2cm 3 EQD2 3

Rectum D2cm 3 EQD2 3

Sigmoid D2cm 3 EQD2 3

Bowel D2cm 3 EQD2 3

OAR

Planning aims

<80 Gy

<65 Gy

<70 Gy

Limits for prescribed dose

<90 Gy

<75 Gy

Preclinical and clinical studies have shown that prolongation of overall treatment time (OTT) is detrimental to patient outcome due to repopulation of tumour clonogens [12]. Analysis of data from the Retro-EMBRACE study [27] suggested that an increase in OTT of 1 week is equivalent to a reduction of CTV-T_HR dose by 5 Gy. In the EMBRACE-II study, it is recommended that the OTT from the first EBRT fraction to the final BT fraction should be ≤50 days. 9.2 Dose constraints The EQD2 concept has allowed the BT equi-effective dose to be combined with the equi-effective dose of different EBRT schedules into a single value which can then be used as a dose constraint for BT dose optimisation. Clinical data based on schedules with a homogenous dose of 45-50Gy whole pelvis EBRT and 40-45Gy of HDR or PDR BT have allowed robust dose constraints for the various target volumes and OAR to be identified [28] and planning aims (as defined in chapter 8.6 of ICRU Report 89) have been introduced. As the target and OAR constraints are often conflicting, hard and soft constraints have been defined for each target volume and OAR to allow for compromises as shown in Table 3. The ability to achieve certain dose constraints depends much on the availability of IC/IS applicators, as IC applicators have significant limitations in large or asymmetrical tumours to reach a high target dose with good OAR sparing. Also, with MRI-based IGABT, target contouring and dose evaluation are more precise and different prescription protocols with different dose planning aims can therefore be applied for different clinical environments e.g. with or without MRI with applicator in place or with or without access to IC/IS applicators. The relative dose contribution from EBRT and BT is also important when selecting EQD2 constraints for certain volumes of OAR, as changing the dose contribution of EBRT compared with BT will significantly influence the dose–effect curves for an OAR. 9.3 Implant geometry The choice of implant geometry is based on tumour, target and OAR topography at the time of BT, individual patient anatomy and institutional practice. The intravaginal component, such as ovoids

Figure 14. Typical median volumes and mean doses for cervical IGABT Unpublished data from EMBRACE I (n=1300) and EMBRACE II (CTV-T LR init , n=168). Initial median GTV-T in EMBRACE II is 55 cm 3 . The anatomical location of the GTV-T init at the time of diagnosis is reflected in the CTV-T_IR defined at BT; this region received a median near minimum dose of 62 Gy in EMBRACE I. In good-responding tumours, the dose at the limits of the GTV-T init is 60-70 Gy, while in poor-responding tumours the region of GTV init may receive doses similar to the CTV-T HR (e.g., around 80 Gy) (Reprinted from Semin. Radiat. Oncol., 29/3, Tan LT, Tanderup K, Kirisits C, et al , Image-guided Adaptive Radiotherapy in Cervical Cancer, 284-298, 2019, with permission from Elsevier).

The whole pelvis is usually treated with EBRT to 45–50 Gy at 1.8-2 Gy/fraction followed by BT to reach total EQD2 values of 85 to 95 Gy for the D90% of the CTV-T_HR (Figure 14) [19]. It is important to avoid a heterogeneous EBRT dose in the central part of the pelvis, e.g., from the use of midline blocks, as this may invalidate the calculation of the combined dose from EBRT and BT particularly to adjacent OAR. The optimal timing of BT in relation to EBRT can be adapted depending on the initial size of the tumour, the response to EBRT and the BT fractionation schedule. For advanced disease and large tumours, significant tumour regression (70-80%) often occurs during the first 3-4 weeks of EBRT and the use of BT after or towards the end of EBRT can greatly improve target coverage and facilitate a response-adapted BT prescription resulting in an improved therapeutic ratio.

Cervix cancer

THE GEC ESTRO HANDBOOK OF BRACHYTHERAPY | Part II Clinical Practice Version 1 - 01/09/2023

and rings, allow laterally asymmetric isodoses to be created in the regions close to the vaginal sources. In contrast, the shape of the isodoses around the tandem become more and more cylindrical superiorly, limiting the possibility of covering laterally situated target volumes. In practice, the planning-aim isodose cannot be placed more than 25 mm from the tandem at the level of Point A; the isodose can be pulled closer to the tandem if the dose to OAR is too high but it cannot be extended much further from the tandem if the target exceeds this distance. Acceptable dose coverage 35 mm from the IU tandem at the level of Point A can be achieved with combined IC/IS applicators with additional lateral needles; these typically cover the medial half of the parametrial space. Applicators with additional obliquely directed needles, or free-hand oblique needle placement, allows lateral dose coverage for distal parametrial and pelvic wall residual disease to be increased (Figure 15). It is important to note that the dose distributions used in published clinical series are still largely dependent on a contribution from the IC application [29, 30]. Needle loading should only be activated within or in immediate proximity to the CTV-T_HR and dwell times per dwell position in needles are normally limited to 10–20% of those used in standard loading patterns for IC implants [31]. 9.4 Pre-planning of the implant Pre-planning according to the definition of the CTV-T HR is necessary to determine the optimal applicator type and to assess the need for an IS approach. As a minimum, a gynaecological

examination should be performed, and the findings recorded on a clinical drawing and compared with the extent of disease at diagnosis. An MRI, with or without an IC applicator in situ, obtained a few days before the actual implant is a useful aid to pre-planning and provides additional information on the response to EBRT and the tumour and OAR topography at the time of BT. Table 4 presents an overview of current possibilities for pre- and intra-operative planning of cervical cancer BT to help make the best individual choice of applicator type, implant technique, and applicator placement in relation to the CTV-T HR and the anatomical situation at the time of BT. 9.5 Applicator reconstruction Recommendations regarding applicator reconstruction for IGABT for cervical cancer were published by GEC-ESTRO in 2010 [10]. Applicator reconstruction defines the relation between the radiation source and the anatomy of the patient in the treatment–planning system, so that the dose contribution from each source position can be calculated for each anatomical voxel. The accuracy of applicator reconstruction is of utmost importance because of the steep dose gradients inherent to BT. Incorrect reconstruction of the applicator will result in dose deviations in both target structures and OAR of up to 5–8% for each mm of applicator displacement [32]. The applicator and source path have to be defined on the individual patient images at the time of BT. This can be done by digitization directly on the acquired images (X-ray, CT or MRI) with the applicator in situ or by importing a library file of the applicator

Figure 15. Combined IC/IS implant with straight and oblique needles (Vienna II applicator) (a) MRI and clinical drawings at the time of diagnosis; (b) assembly of applicator; (c) MRI and clinical drawings at the time of brachytherapy with Vienna II applicator to encompass the residual disease in distal parametrium on right; (d) contouring the target and the organs at risk; (e) BT treatment planning showing 3D view of the applicator with straight and oblique needle reconstruction and final plan showing isodose distribution in different views (Reprinted from Radiother. Oncol., 141, Mahantshetty U, Sturdza A, Naga Ch P, et al, Vienna-II ring applicator for distal parametrial/pelvic wall disease in cervical cancer brachytherapy: An experience from two institutions: Clinical feasibility and outcome, 123-129, 2019, with permission from Elsevier).

Cervix cancer

THE GEC ESTRO HANDBOOK OF BRACHYTHERAPY | Part II Clinical Practice Version 1 - 01/09/2023

TABLE 4 STRATEGIES FOR PRE-PLANNING Adapted from Table 12.1, ICRU Report 89

When contouring and reconstruction are performed in the same image series, the dose to targets and organs can be directly calculated without any image fusion. However, there are several circumstances when contouring is done in one image series or modality and reconstruction in another. When different image sets are used for contouring and reconstruction, they have to be co-registered according to the applicator [34]. Fusion uncertainties will translate into dose calculation uncertainties and can result in a miscalculation of dose-volume histogram (DVH) dose parameters by 4-6% per mm of fusion error [35]. Matching to bony structures must be avoided as the applicator can move more than 5 mm in relation to bone [36]. For X-ray based treatment planning, correct definition of reference points for target and OAR dose reporting is key (see 10.1.1). With volumetric MRI or CT-based BT, although dose prescription is mostly linked to target volumes and no longer to Point A, the reporting of point doses is still recommended and relevant. The definitions of these reference points on 3D images are described in chapter 9.2 of ICRU Report 89. 9.6 Loading patterns and optimisation of the dose distribution Remote afterloading equipment with stepping source technology typically has active dwell positions in step sizes of 1 mm to 10 mm. The width, height, and thickness of the isodose surface volumes (ISV) for the same BT applicator can differ substantially based on the loading pattern. In one study, seven different European and American institutions provided their standard loading patterns for a tandem-ring applicator and tandem-ovoid applicator [37]. The width of the normalised Point A ISV reported in these studies ranged from 4.8 cm to 6.0 cm at the level of the ring applicators. Standard sets of active dwell positions which simulate standardised radium-tube configurations can and have been used. These applicator-based loading patterns often result in unfavourable dose distributions in non-ideal applicator placements with asymmetrical tumour or OAR positions. In general, when using standardised Clinical examination will inform decisions on applicator type and implant geometry. This is used in the “mould technique,” where the applicator is modelled from a vaginal impression. An MRI just prior to brachytherapy allows assessment of the configuration and dimensions of the CTV_HR and the topography of OARs. Note that when the intracavitary applicator is inserted, the topography of the uterus and target volumes and their relation to the OAR will change substantially. The intracavitary applicator is inserted to serve as a template for optimal needle insertion which can be used to generate a full treatment pre-plan with regard to target coverage and OAR constraints. Before insertion of the brachytherapy applicator, a comprehensive clinical examination is essential to plan the implant. The insertion of the intrauterine tandem and parametrial needles can be guided by trans-abdominal or trans-rectal US Post-implant imaging allows for evaluating the quality of an implant with regard to dose–volume constraints for target and OAR.

Without applicator in place

Without volumetric imaging

With applicator in place

Pre-operative

Without applicator in place

With volumetric imaging

With applicator in place

Without volumetric imaging

During insertion of the applicator

Intra-operative

With volumetric imaging

With applicator in place

geometry. Direct digitization can be used when the source channels or marker wires are visible in the images. A library of applicators can be defined for fixed-geometry applicators and merged with the patient images based on anchor points or by direct positioning of the applicator shape according to visible structures of the applicator on the images (e.g. source path or outer surface of the applicator). With CT images, it is possible to directly visualise the source channel either by exploiting the contrast between the applicator material and the air-filled source channel or by inserting wires with radio-opaque markers separated by fixed gaps. Marker strings may have different flexibility and dimensions from the source wire and discrepancies in source position of 2–3 mm have been observed [10, 33]. Reconstruction of applicators with MRI is more challenging than with CT as the source channel is not visible due to lack of MR signal from air and the applicator materials. Furthermore, markers used for radiographs and CT cannot be used for MRI. Special MR markers, such as catheters containing a CuSO4 solution, water, glycerine, or US gel, can be inserted into the source channels in plastic applicators. Reference structures, such as needle holes or cavities filled with fluid, can also be used as long as the locations relative to the dwell positions are known. Fluid-filled marker catheters cannot be visualised inside the source channels of titanium applicators and there are also uncertainties due to distortions and artifacts that can change with MRI pulse sequence. MRI slice thickness has direct impact on the precision of reconstruction, and it is recommended that reconstruction be performed in an image series obtained with a slice thickness of ≤5 mm. Reconstruction of plastic needles on MR images can be particularly challenging as there are no commercially available MR needle markers and reconstruction is entirely reliant on image quality. Knowledge of needle insertion depth is crucial for correct needle reconstruction and can be obtained by measuring the outside needle length from the vaginal applicator or template.

Page 2 Page 3 Page 4 Page 5 Page 6 Page 7 Page 8 Page 9 Page 10 Page 11 Page 12 Page 13 Page 14 Page 15 Page 16 Page 17 Page 18 Page 19 Page 20 Page 21 Page 22 Page 23 Page 24 Page 25 Page 26 Page 27 Page 28 Page 29 Page 30 Page 31 Page 32 Page 33 Page 34

Made with FlippingBook flipbook maker