Intersex and cancer
Some intersex conditions present an elevated risk of gonadal cancer
(cancer in the tissue that forms testes or ovaries). There are different kinds
of cancer, and the main risk in intersex people is presented by what are called
type-II germ cell tumours (GCT).
Removal of gonadal tissue, called gonadectomy, may be an appropriate treatment
to manage the risk. The cancer risks however are complex and in some cases
poorly understood. The chance that gonads will develop cancer depends on
several variables, and the stage in life at which the risk becomes significant
varies depending on the type of intersex (though there is limited information
available about this).
Removal of gonads in infancy is sometimes recommended in order to nullify the
risk of cancer, but this potentially conflicts with the principle of avoiding
irreversible surgery on a child unless necessary, to allow the person an
opportunity to make their own decisions regarding their medical treatment.
In some cases retention of gonads is also desirable to preserve natural hormone
It was reported to the committee that there is a trend toward fewer removals
of gonads during infancy as a result of changed approaches to intersex.
This is a positive development, but it does mean that more attention now needs
to be paid to the health risks – particularly the tumour risk – that may arise
from those gonads being retained.
As was noted in chapter 1, some intersex people are fertile, and others
are not. It depends on the type of intersex variation a person has, as well as on
the specifics of their case. Removal of gonads in many cases would not be
sterilising, because they would not be fertile in the first place. In some
cases, however, gonads may be fertile, or may contain tissue that could allow
fertility as a result of future advances in medicine. As some decisions to
remove gonads are made shortly after birth, this means removal occurs 20 to 40
years before the person might seek to have children – a very long period over
which to predict what advances in medicine might occur.
It is certainly the practice of some specialists to avoid removing
potentially fertile tissue wherever possible.
This reflects the 2006 Consensus Statement position that 'surgical management
of DSD should also consider options that will facilitate the chances of
The main issues raised during the inquiry concerned the estimation of
cancer risk, and the way in which medical intervention relies on assessment of
those risks. Because there was disagreement amongst participants in the
committee's inquiry regarding the levels of cancer risk in intersex people, and
appropriate medical responses to those risks, the committee considered in more
detail the published research that lies behind this discussion. This chapter
explains how cancer risks and diagnostic techniques have been set out in the
medical literature. An important part of this discussion involves a table of
information that is reproduced in different forms in many publications, and
which appeared to be the source of some of the problems that have emerged
during debate about gonadectomies in intersex people.
Reviews and clinical recommendations in the medical literature
Beginning in 2005, a team of researchers largely based in Rotterdam in
the Netherlands published a series of articles and reviews regarding the
nature, diagnosis and treatment of germ cell tumours in intersex people.
Throughout the literature, the discussion is of different 'disorders of sexual
development' (DSD), and for consistency of reference to the literature, that
terminology will frequently be used in this section of the report. The work of the
Dutch team has been pivotal in improving the understanding and management of
gonadal cancer in intersex people.
In 2006, the group published a key review of evidence, titled 'Germ cell
tumors in the intersex gonad: old paths, new directions, moving frontiers'.
This paper assembled evidence from over a hundred studies in the field, and
made a number of contributions, including:
proposal that both the classification and terminology associated with DSD be
synthesis of data, leading to a summary of 'the estimated germ cell tumor
prevalence in patients with DSD', according to the type of DSD;
new test and diagnostic approach, to reduce over-diagnosis of cancer or cancer
risk in some patients;
new classification model for patient risk, based on morphology and histology;
table ('the 2006 table') that set out a 'summary of the risk of germ cell
malignancy in the various forms of DSD, subdivided into high, intermediate, low
and possibly no risk' including a column describing 'action needed'.
It was this table and its subsequent incarnations that featured
regularly thereafter in publications, including in submissions to the current
inquiry. The 2006 table is reproduced in full, below:
TABLE 7. Summary
of the risk of germ cell malignancy in the various forms of DSD, subdivided
into high, intermediate, low, and possibly no risk
No. of studies
No. of patients
Watchful waiting and possible
Biopsy4 and possible
Testicular tissue removal in
case of ♀ rearing?
Leydig cell hypoplasia
GD (+Y)2 scrotal
Biopsy4 and irrad?
PAIS scrotal gonad
Biopsy4 and irrad?
actions are indicated, as well as the number of studies and patients included
in the survey. In case of PAIS, 17β-HSD, and ovotestis, the decision
regarding gonadectomy is largely determined by sex of rearing. Relevant data
from the recently published study by Hannema et al. (107 ) are not
included in this table because it is at present unclear to us to what extent
patient series from this study show overlap with patient series from a
previously published study by the same group (117 ). Intra-abd, Intraabdominal
located gonad; nonscrotal, nonscrotally located gonad; scrotal, scrotally
located gonad; irrad, local irradiation with 18 Gy; ovotest. DSD, formally
ovotestis (true hermaphrodite).
- GD (including not further specified, 46XY, 46X/46XY, mixed, partial, complete).
- GBY region positive, including the TSPY gene.
- At time of diagnosis.
- At puberty, allowing investigation of at least 30 seminiferous tubules,
preferential diagnosis based on OCT3/4 immunohistochemistry.
- PCR detection of Y-chromosomal sequences (in particular the GBY region) is
implicated if a marker is identified by karyotyping.
- Based on current knowledge (single study including limited number of patients,
or no studies reported at all).
A version of this table also appeared in the 2006 'Consensus Statement
on management of intersex disorders' (discussed in chapter 2). There are
however some significant differences between the versions of the table.
Compared to the table above, in the 2006 Consensus Statement version of the
of the disorders – gonadal dysgenesis (GBY region positive, including the TSPY
gene) with scrotally located gonad; and partial androgen insensitivity syndrome
with scrotally located gonad – have been moved from the 'unknown' or 'possibly
no' risk category, and placed in the 'intermediate risk' group;
'unknown' risk category is re-titled 'no(?)' risk;
the case of partial androgen insensitivity syndrome (PAIS) with non-scrotally
located gonad, it appears one of the three studies used to provide an estimate
has been omitted, reducing the number of patients from 80 to 24, and
significantly increasing the estimated risk, from 15 per cent to 50 per cent;
the case of complete androgen insensitivity syndrome (CAIS), it also appears
one of the three studies used to provide an estimate has been omitted, reducing
the number of patients from 120 to 55, and increasing the estimated risk, from
0.8 per cent to 2 per cent;
main footnote, with its explanation for recommended action in the case of PAIS,
17β-HSD, and ovotestis, is omitted;
proposed action for CAIS is changed from 'Biopsy and possible irrad/gonadectomy'
to 'Biopsy and ???';
proposed action for Ovotesticular DSD is changed from 'Testicular tissue
removal in case of ♀ rearing?' to ' Testicular tissue removal?'; and
last two footnotes are omitted.
There were two main effects of these changes. First, they increased the apparent
level of cancer risk of some intersex conditions. Second, they removed explanation
of the table's content regarding links between the preferred course of action and
the chosen sex of rearing, but without removing or modifying the courses of
action based on those explanations.
A 2007 paper by the research team, titled 'Tumor risk in disorders of
contained a table substantively identical to that in the 2006 consensus
statement. The table (as reproduced in one of the committee's submissions) is
Source: APEG, Submission
88, citing Leendert H.J. Looijenga, Remko Hersmus, J. Walter Oosterhuis,
Martine Cools, Stenvert L.S. Drop and Katja P. Wolffenbuttel, 'Tumor risk in
disorders of sex development (DSD)', Best Practice & Research Clinical
Endocrinology & Metabolism, Vol. 21, No. 3, 2007, p. 491.
The 2006 and 2007 papers from the Dutch research team, and the 2006
Consensus Statement, contain no information regarding the omission of one study
from the sample, or accounting for the other changes.
The team of researchers responsible for the 2007 paper were cautious in
their presentation of the information. They stated that the application of a
combination of diagnostic techniques presented by them 'might in future be
used to develop a decision tree for optimal management of patients with DSD'
The authors concluded that 'patients with DSD can be classified into high,
intermediate, low or unknown risk groups for type-II germ-cell tumors'.
They qualified the classification, however, by noting that there are some
intersex conditions for which no or insufficient data is available (including 5α-reductase
deficiency and Leydig-cell hypoplasia), and by indicating that:
This first attempt to estimate the risk of the individual
patient with DSD developing a type-II germ-cell tumor must be tested using
additional cases in which proper criteria are used for classifying patients in
the different DSD entities...
In 2009, researchers Professor Gary Warne and Dr Jacqueline Hewitt
published a paper in the Medical Journal of Australia, titled 'Disorders of sex
development: current understanding and continuing controversy'. Based on the
2007 results in Looijenga et al, Warne and Hewitt stated, regarding risks of
In any DSD associated with a Y chromosome, there is an
increased risk of germ cell cancer, especially when the testes are
intraabdominal (the risk of seminoma in partial androgen insensitivity is 50%
for an intra-abdominal testis) or when there is gonadal dysgenesis.
In relation to clinical management of those children in whom testes are
retained, Warne and Hewitt continued:
The trend for surgeons to recommend male-sex rearing for
greater numbers of children with DSD could also mean greater reluctance to
remove testes that pose a significant risk of cancer on the grounds that
physiologically useful hormone secretion might be retained. It is therefore
imperative that a risk management strategy be prepared for each patient. This
- educating parents and patients about risk;
- removing all intra-abdominal gonads that cannot be brought
down into the scrotum;
- regular clinical and ultrasound surveillance of scrotal
gonads with removal of any that contain suspicious lumps;
- biopsy of testes after the onset of puberty, looking for early
signs of malignant change; and
- effective communication between paediatric and adult care
providers at the time of transition.
In 2009, the Dutch team published another review paper that incorporated
a table similar to that published in 2006, with the larger number of studies
for CAIS and PAIS (3 in each case rather than 2), and repeating the lower risk
estimates of their 2006 publication.
In 2010, the Dutch research team (in a publication with a Czech lead
author, J. Pleskacova) published a further paper in the field, somewhat
confusingly carrying the same title as the 2007 article. The 2010 publication
did not carry a version of the same table, but did contain a similar, smaller
table summarising the prevalence of GCT in DSD patients. This table was as
Notably, the estimates of prevalence for two key disorders, PAIS and
CAIS, reflect again the lower estimates in the 2006 publication, which is cited
as the principal source for the table.
The team of researchers concluded that:
Presently available tools allow us to assess gonadal tissue
of DSD patients and identify gonads at risk for GCT development, i.e. gonads
containing dysplastic cells or noninvasive neoplasia. This ability together
with precise diagnosis of DSD cases based on molecular-genetic methods may
facilitate a more accurate estimation of the tumor risk in various forms of
DSD. With that knowledge we might be able to preserve gonads in selected
The Dutch team published two papers in 2011. Neither included a table of
data equivalent to that found in earlier publications. One of the 2011 papers
related only to one subset of DSD: 45,X / 46,XY mosaicism.
The other focussed on 'tumor risk in relation to the gonadal differentiation
pattern and the phenotypic presentation of the patient'.
The papers do not directly discuss the risks associated with CAIS or PAIS, but
do indicate the developing understanding of the relationship between tumour
risk and the location and nature of gonadal tissue, concluding 'tumor risk is
most pronounced in immature and/or poorly differentiated gonadal tissue and can
be – at least in part – predicted from the presence of specific
Discussion during the committee inquiry of the medical research
APEG made a submission to the committee's inquiry that reproduced the
table from the 2007 paper, described above. APEG's position in its submission
In high-risk groups the recommendation is to remove the
gonads before the individual develops cancer, which can occur in childhood. It
would be negligent to expose these children to cancer by leaving the
testes/ovaries in when the high risk is known...[and] The recommendation of Warne
and Hewitt, and in the current medical literature, is for preventative surgical
removal only in the high-risk and intermediate-risk cancer group...
The Disorder of Sex Development multidisciplinary team at Royal
Children's Hospital, Melbourne (RCH) also discussed the risk of cancer. It reported
a number of figures, including the 50 per cent figure that appeared in Warne
and Hewitt's 2009 paper, again citing the Dutch team's 2007 publication as the
XY Complete gonadal dysgenesis. Individuals with this
condition may have both the external physical appearances of a girl and a
uterus, and will most likely identify as female. If [their] gonads are
intra-abdominal, there is 15-30% risk of malignancy occurring by the time the
young woman reaches her mid 20’s...
Partial androgen insensitivity syndrome (PAIS)...There is a
considerable spectrum – with some people being born with almost normal male
external genitalia, and others having almost normal female genitalia (but all
will have no uterus). If the testes are undescended and inside the abdomen, the
cancer risk of the testes is reported to be 50%.
It should be noted that the 2007 paper cited by the Melbourne team
claimed only to summarise major findings from other papers, and referred the
reader to the 2006 paper and the Consensus Statement for details.
Of these, only the 2006 paper discusses risk for individual types of intersex
in detail, and gives a reported cancer risk for PAIS of 15 per cent rather than
The 15 per cent figure is likewise reproduced in 2009 and 2010 papers from the
In its submission, OII commented on the 2009 Warne and Hewitt paper. OII
Warne and Hewitt’s assertion regarding the percentage risk of
malignancy in internal gonads strongly imply a general, across the board, risk
of 50%. This is considerably different from research elsewhere, suggesting
either sampling bias, or a hitherto unknown cancer hot spot...
The protocol described by Warne and Hewitt means that the
testes of all people with CAIS, and very many with PAIS, are removed in
infancy. Alternative views are numerous, including international expert Katrina
Karkazis or, in the case of AIS specifically, by Quigley et al Batch et
al, Crouch. The AISSG UK summarise some of the research in this field,
showing sampling bias in many studies, and far lower risks for most intersex
people with internal gonads, albeit risks that increase with age.
OII's submission went on to cite some of that research, which gives
differing rates of cancer risk for different types of intersex condition.
The two studies mentioned by OII that were published in medical journals
(Quigley et al and Batch et al) pre-date the work of the Dutch
team (and others), who were able to draw on new diagnostic techniques and
larger sample sizes.
Professor Warne and Dr Hewitt were co-authors (with others) of the APEG
submission. That submission responded to evidence from OII, stating:
The Senate has unfortunately received misleading information
in submissions on this issue. We are concerned that some of the information
presented appears to have been either misunderstood, or misrepresented in
error, leading to inaccurate conclusions. Some authors have misunderstood the
difference between high-risk and low-risk cancer groups within DSD, and in
particular, one submission incorrectly implied that the cancer risk for a
diagnosis in the highest-risk group ('PAIS with non-scrotal/intra-abdominal
testes') was quoted by Warne and Hewitt as being the cancer risk for a diagnosis
in the low-risk group ('CAIS'), as outlined in Table 2. The implication is that
testes or ovaries are being removed from patients with diagnoses at low-risk of
cancer, such as CAIS, however this is incorrect.
Subsequent submissions appear to indicate that there is some common
in recognising that cancer risk in some intersex people, especially those with
CAIS or ovotesticular DSD, does not warrant prophylactic removal of testes.
At the same time, OII, quoting other medical research,
maintained that testes are still being removed from low-risk individuals
(though presumably not by those specialists who do not support the practice,
such as Warne and Hewitt, or the team at RCH Melbourne).
The committee received no evidence on the numbers of gonadectomies being
performed where surgery was based on cancer risk.
The committee wrote to authors of the published research, seeking
clarification of the variation in the estimated cancer risk or prevalence
between different studies. In responding, the group of medical experts noted:
In any individual with a DSD condition, the decision to
perform gonadectomy is reached by weighing benefits and risks of various
issues, such as risk for [germ cell tumour], sex of rearing, estimated capacity
of the gonad to produce hormones in accordance with or opposite to sex of
rearing and/or (developing) gender identity, likelihood of gender dysphoria
later in life, etc.
The statement 'In case of PAIS, 17β-HSD, and ovotestis,
the decision regarding gonadectomy is largely determined by sex of rearing'
should be interpreted in this broader and clinically oriented context, which is
different from the studies presented later, focusing primarily on tumor risk
and in which the clinical emphasis is less elaborated.
The committee identified two related issues in the discussion of
intersex and cancer risk:
complexity and diversity of cancer risk can become oversimplified, potentially elevating
the perceived or communicated risk. Alternative monitoring options may be
committee is concerned that other matters such as 'sex of rearing' or
'likelihood of gender dysphoria' are interpolated into the discussion of cancer
risk. This confusion between treatment options to manage cancer risk and
treatment options to manage intersex could undermine confidence in the
neutrality of those advocating for surgical interventions.
One of the difficulties faced by the committee and others when
considering this literature is that the application of labels such as 'low
risk' or 'high risk' appears to be masking some of the variation between
individual intersex conditions. There are also serious questions to be raised
about what constitutes 'high risk', and why it is that cases facing an
'intermediate risk' should be subject to prophylactic gonadectomy in infancy.
As cited above, the APEG submission stated that 'The recommendation of
Warne and Hewitt, and in the current medical literature, is for preventative
surgical removal only in the high-risk and intermediate-risk cancer group'. However
the detail is more complex. In intermediate risk cases, the published
literature has recommended gonadectomy only in some cases. For others, there is
no definite recommendation.
The summary classification of intersex conditions by cancer risk may also
mask the importance of considering the circumstances of individual cases. There
is great genotypic and phenotypic diversity among intersex people, even within
a single category of intersex condition, and the literature suggests that these
specific circumstances have a bearing on the cancer risk. As Dr Cools pointed
The risk of GCT development varies undoubtedly according to
which DSD a person has. However, in view of the very low incidence of most DSD
conditions, and given the fact that gonadectomy has been performed
prophylactically at an early age in many cases, it is currently impossible to
obtain correct estimates of this risk for every DSD condition... any statement
about tumor risk on an individual basis is an estimate and is possible
only after thorough diagnostic investigations, most often including gonadal
biopsy taking and specialized immunohistochemical
analysis, which needs expert surgical manipulation and centralization of
material, with specialist analysis.
There are, for example, some types of intersex that are generally classed
as at high risk of gonadal cancer, but in which the published research papers
indicate that the risk of tumour development depends on the morphology and histology
in the individual case.
These include people with dysgenetic testes or with 'undervirilising'
conditions such as PAIS. For these intersex people, and others, a number of
specific factors can be examined in the individual that will influence whether
they or not they are at high risk of developing tumours.
As the Dutch team concluded in one of its most recent papers:
Tumor risk is most pronounced in immature and/or poorly
differentiated gonadal tissue and can be – at least in part – predicted from
the presence of specific immunohistochemical markers. This increase in
knowledge has modified our clinical approach to the DSD patient, resulting in
an individualized management with regard to tumor risk.
It is also the case that the authors of the published research continue
to repeat their cautions that the estimates and diagnostic models are only
preliminary, and are in need of further empirical validation.
In these circumstances, the quoting of some of the risk estimates, particularly
the higher ones relating to PAIS and 17β-HSD, appears not necessarily to
be based on strong evidence. Quoting some of these summary estimates has the
potential to hinder the process of objectively assessing individual patient
risk, and of ensuring that cancer-related treatment considerations are kept
visibly separate from other factors (such as urogenital corrective surgery, or
Action to manage cancer versus
action to manage intersex
The committee concluded that one of the causes of disquiet regarding the
management of cancer risk is that some of the published literature does not
adequately distinguish between the appropriate clinical course of action
regarding an intersex person's risk of cancer, and the appropriate clinical
course of action to manage a person's intersex condition itself.
The footnotes to the 2006 table, missing from other later versions, encapsulate
this problem. The 2006 table had notes making it clear that the recommended
actions did not arise solely from the cancer risk associated with a variety of
intersex, but took account of other factors such as the proposed sex of rearing
of the child. This clarity was lost once such notes were omitted. By far the
most serious omission was in the case of the 2006 Consensus Statement, because
of its broad scope and considerable influence.
In answering the committee's questions about the communication of cancer
risk in the literature, Dr Cools and others argued that the 2006 paper
discusses the risks in a 'broader context' that is 'different from the studies
presented later, focusing primarily on tumor risk'.
The balance of evidence does not support this. The paper that they describe as
having a 'broader context' is specifically titled 'germ cell tumours in the
intersex gonad', and its abstract refers solely to tumour risk and developments
in the field in relation to this. The one apparent exception within that paper
– the authors' development of an alternative classification schema for intersex
– is itself 'proposed as a tool to refine our insight in the prevalence of
germ cell tumors in specific diagnostic groups'.
Within that paper the table is headed 'summary of the risk of germ cell
malignancy in the various forms of DSD, subdivided into high, intermediate,
low, and possibly no risk', contradicting the argument that the paper is discussing
treatment in a broader context compared to later papers, where the table has a
very similar title in all cases. Furthermore, if this explanation was correct,
then the 2006 Consensus Statement – which definitely does have a far broader
context than any of the individual research papers – should have the additional
explanations included, yet it does not. Furthermore, the 2006 Consensus
Statement explicitly describes the 'recommended actions' in the table as being
'recommendations for management' of the risk of tumour development, not
management of the person's intersex condition generally.
This does not appear consistent with the explanations offered in the 2006 paper
and elsewhere. Finally, the experts' answer to the committee's question argued
that the explanations are included in the 2006 paper because of its 'clinically
oriented context', yet the 2007 paper, which lacks the explanations, is even
more explicitly clinically oriented, appearing in the journal 'Best Practice
and Research Clinical Endocrinology and Metabolism', where each section of the
article concludes with 'practice points' for clinicians.
Dr Cools and her Dutch team have sought to advance the scientific
understanding and estimation of cancer risk in intersex individuals, and have done
so with considerable success. This was intended to provide better information
about one key factor in intersex medical decision-making (assessing the
patient's cancer risk). Instead, because of the incorporation of a table column
listing 'recommended actions' based in part on consideration of other factors
such as sex of rearing but with that explanation frequently omitted (most importantly
from the 2006 Consensus Statement), the information risks being interpreted as
a guide to clinical action on the grounds of cancer risk, which it is
This detail is important. There is considerable debate, some of it
outlined in the previous chapter, about the merits of performing surgery at
different ages. Intersex organisations, regulators, courts and other
decision-makers are closely scrutinising, and sometimes relying on, this
medical literature to inform this extremely important discussion taking place
in the broader community, beyond just the medical professions. To allow this
debate to take place transparently and with the confidence of the intersex
community, it is essential that the different reasons for medical treatment,
and the attendant risks, are characterised separately. Otherwise, decision-making
becomes opaque to families, courts, regulators, support groups, and even to
external clinicians. This will undermine confidence, in turn prompting calls for
blanket bans on particular medical procedures, removing clinicians from decision-making
processes. The committee would see these as undesirable outcomes.
The committee is aware of a risk, not directly discussed by witnesses to
the inquiry, that clinical intervention pathways stated to be based on
probabilities of cancer risk may be encapsulating treatment decisions based on
other factors, such as the desire to conduct normalising surgery. This kind of
encapsulation of factors under a single reason is evident in the published
tables discussed in this chapter. This might happen because of the distinction
made by Australian courts between 'therapeutic' and 'non-therapeutic' medical
intervention. Treating cancer may be regarded as unambiguously therapeutic
treatment, while normalising surgery may not. Thus basing a decision on cancer
risk might avoid the need for court oversight in a way that a decision based on
other factors might not. The committee is disturbed by the possible
implications of this.
If the distinction between therapeutic and non-therapeutic treatment
were to be retained, then the committee would draw attention to an example used
in Queensland legislation relating to guardianship and the circumstances in
which a court should be involved in decisions. The example suggests that
decision-makers may need to distinguish between treatment of cancer, and
treatment for the possible risk of cancer.
If the child has cancer affecting the reproductive
system and, without the health care, the cancer is likely to cause serious or
irreversible damage to the child’s physical health, the health care is not
An implication of the example is that a treatment for the risk of
cancer may not fall into the category of treatments that do not require
The committee shares others' concerns, however, outlined in the next
chapter, regarding the current way in which decision-making occurs for intersex
people who are unable to make the decisions for themselves (generally
children), including concerns about the distinction between therapeutic and
non-therapeutic treatment. The committee does not favour the status quo.
Chapter 3 and this chapter have both outlined how complex and contentious is
some of the information that supports medical treatment of intersex people.
In light of the complex and contentious nature of the medical treatment
of intersex people who are unable to make decisions for their own treatment,
the committee recommends that oversight of these decisions is required.
The next chapter considers what such a system of oversight should look
Navigation: Previous Page | Contents | Next Page