M a r k E m b e... L o u i s D e n... J o h n F i t z...

Fitzpatrick_edit.qxp
25/9/08
10:49 am
Page 16
Prostate Cancer
HistoScanning™ and Its Role in Prostate Cancer Diagnosis,
Staging, Treatment and Monitoring
a report by
M a r k E m b e r t o n , 1 L o u i s D e n i s 2 and J o h n F i t z p a t r i c k 3
1. Reader, Interventional Oncology, University College Hospital, London; 2. Director, Oncology Centre Antwerp;
3. Chairman, Department of Surgery, Mater Misericordiae Hospital, Dublin
During the European Association of Urology (EAU) congress in Milan in
risk disease and as a result will be offered active surveillance as the
March 2008, 27 leading European prostate cancer experts convened
preferred disease management modality.1 These men will be monitored
under the guidance of the authors to discuss the practical implications of
closely and definitive treatment offered only if signs of progression are
HistoScanning™, a new technology to improve cancer diagnosis by
detected. Currently, this process relies on both PSA kinetics and repeat
non-invasive imaging tools. The participants reviewed the non-invasive
biopsy.2 The former has yet to be validated in this context, and the latter
diagnostic methods that are currently available for prostate cancer,
is invasive and suffers from sampling artefacts. If a test was available that
considered current and future needs and tried to assess how well new
could reliably detect any increase in the volume of the tumour under
technologies such as prostate HistoScanning may meet those needs.
surveillance, it is likely that the active surveillance option would be offered
to patients more frequently, especially if this test was non-invasive, safe
The Clinical Need for Non-Invasive
and reproducible. One other development needs to be mentioned in this
Diagnostic Technologies
context. This relates to the idea that there may be a role for focal (gland-
Over the past 10 years there has been an increase in patient awareness of
preserving) treatment in some men diagnosed with prostate cancer.3
prostate cancer, and the resulting prostate-specific antigen (PSA) testing in
Although it is currently an investigational approach, focal therapy is an
the primary care setting has led to a significant rise in the number of men
innovation that makes exacting requirements on the characterisation of
presenting with elevated PSA. Consequently, there has been an increase
the disease in terms of burden, orientation and specifying which parts of
in the number of cancers detected at an early stage, but this has come at
the prostate are to be treated. This requirement for precision is likely to fall
the cost of many men undergoing unnecessary prostate biopsies. A
on imaging platforms that can localise disease accurately.
relatively inexpensive, safe and non-invasive test that could reliably rule
out clinically important prostate cancer would help to mitigate this
Thus, today there is a clear need for a simple and relatively inexpensive
problem. Earlier biopsies have also resulted in an increase in the
non-invasive diagnostic test with acceptably low false-positive and false-
proportion of men who are diagnosed with disease that has a low
negative rates that can provide diagnostic support for the new therapy
probability of leading to a prostate-cancer-related death. This means
regimes and assist urologists in:
that a significant proportion of patients will be deemed to have low• ruling out cancer;
Mark Emberton is a Reader in Interventional Oncology at
University College London and an Associate Professor at
Middlesex University. He is also Clinical Director of the Clinical
Effectiveness Unit at the Royal College of Surgeons of England.
He is involved in guideline production and formulation of
evidence-based management policies. Mr Emberton is an
active researcher, has lectured widely and has published over
100 articles in numerous peer-reviewed journals.
E: [email protected]
• identifying differentiated tissue before biopsy to enhance yield;
• improving risk stratification (lower the risk of under- or over-staging);
• planning treatment more reliably and effectively;
• evaluating changes in the lesion over time (progression/regression); and
• guiding treatment through visualisation.
How Good Is Current Imaging?
Over the past few decades, the development of transrectal ultrasound
has assisted in diagnosing prostate cancer. Improved biopsy methods
Louis Denis is Director of the Oncology Centre Antwerp,
Treasurer of the International Consultation on Urological
Diseases (ICUD) and Chairman of the International Prostate
Health Council (IPHC). He was a Co-ordinator of the European
Randomised Screening Study for Prostate Cancer and Founder
of the European Coalition against Prostate Cancer (Europa
Uomo). He has edited or co-edited over 50 books and
published more than 300 peer-reviewed articles.
make it possible to take samples that are spaced around the prostate
when there is some indication of malignancy, such as a high PSA level.
Various types of transrectal ultrasound are now routinely used and other
imaging technologies are also available, but not one of the current
methods meets all of the clinical needs, and all have limitations (see Table
14–21). Magnetic resonance imaging (MRI) comes closest to providing the
necessary information, but its cost rules out its use as a widespread
John Fitzpatrick is a Professor and Chairman of the
Department of Surgery in the Mater Misericordiae Hospital and
University College Dublin. He is on the Editorial Board of 25
journals and is Editor in Chief of BJU International. He is Past
President of the British Association of Urological Surgeons and
the Irish Society of Urology, and a Member of the Board of
Trustees of the British Urological Foundation.
diagnostic method. Ultrasound meets the accessibility and affordability
criteria, but the detail that is visible to the user on conventional
ultrasound displays is somewhat limited.
Making Use of Ultrasound’s ‘Lost’ Data
Modern ultrasound equipment can detect structures in the range of
10-5m. By the time the data captured have had the necessary
16
© TOUCH BRIEFINGS 2008
Fitzpatrick_edit.qxp
25/9/08
10:50 am
Page 17
HistoScanning™ and Its Role in Prostate Cancer Diagnosis, Staging, Treatment and Monitoring
Table 1: The Usefulness of Various Imaging Technologies for
Prostate Cancer Diagnosis, Staging and Monitoring Treatment
Technology
TRUS
Benefits
• Easy, available, quick, safe
and cheap.
• Good visualisation
of palpable tumours
• Useful for guiding biopsies
and prostate volume studies4
Colour and power
Doppler US
•
•
Contrast-enhanced
US
Sonoelastography
•
•
•
Computed
tomography
MRI and MRSI
•
•
•
•
•
•
Positron emission
tomography
•
Limitations
• Variable specificity
(50–90%)5,6 and
sensitivity (48–86%)6
• 30% palpable tumours
not visualised
• Rarely detects small
(<5mm) lesions
• Low local staging capability;
unable to show EPE7
Specificity better than
• Sensitivity only a little
greyscale US for lesions
better than greyscale US
>1cm in transition zone
• Cannot detect small
Relatively accessible
lesions (<5mm) with
and affordable
micro-neovascularity8
May improve detection of
• Inherent weak signals
smaller (2–5mm) lesions
difficult to detect from the
background signals
• Expensive, time-consuming,
complicated9
• May require pre-medication
to reduce blood flow10
Yields a colour mapping
• Relatively low sensitivity
of tissue elasticity
(68%) and specificity (81%)11
Relatively affordable
• Examiner-dependent12
• Poor at detecting posterior
tumours12
• Narrow dynamic elasticity
range of the prostate12
Often used to look for
• Lacks soft-tissue
lymph node involvement
contrast resolution
• Unable to distinguish
malignant from
non-malignant zones13
• Relatively expensive and not
universally available
Accurate for staging disease
• Low sensitivity14
High specificity14
• BPH may obscure lesions
Able to detect EPE, seminal
in the transition zone17
vesicle, bladder or rectum
• Expensive and therefore
not universally available
invasion14,15
Useful for treatment planning • MRSI (citrate reduction) may
MRSI increases staging accuracy
confuse prostatitis or
and reduces interobserver
post-biopsy haemorrhage
variability16
with cancer lesions18
Can detect metastatic
• Uptake of radiotracer
disease
18-FDG may vary when
cancer present19
• Inability to distinguish scar
tissue from local recurrence
post-prostatectomy20
• At best provides similar
information to TRUS or MRI21
• Expensive and not
widely available
TRUS = transrectal ultrasound; EPE = extraprostatic extension; US = ultrasound;
CT = computed tomography; MRI = magnetic resonance imaging; MRSI = magnetic resonance
spectroscopy imaging; BPH = benign prostatic hyperplasia.
Figure 1: Identifying Differentiated Tissue
The large background region (A) is composed of spotted balls, but the coloured area is
composed of dice, and therefore differentiated from the background. However, the two areas
cannot be distinguished visually. Differentiating the group of dice is not possible even if the
picture is enlarged, because information has already been lost in the display process. The
original data captured to create this picture contain all of the necessary information needed for
differentiation, such as different reflection characteristics between spheres and cubes and the
resolution required to recognise the different patterns. Mathematical and statistical analysis of
all of the original data captured would reveal the difference between areas B and C. Similarly,
HistoScanning analyses all of the source data underlying the digital ultrasound image and
would reveal the differentiated area in the ultrasound image, allowing it to be highlighted.
detection, characterisation and visualisation of differentiated tissue,
such as prostate cancer lesions. For any tissue characterisation
technology to deliver consistent and optimal results it is important that:
• data collected are standardised and user-independent; and
• early source data are used for analysis. (Data compression, dynamic
range mapping and filtering all result in the irreversible loss of
important information.)
For this reason, the latest version of the HistoScanning technology uses
radiofrequency (RF) ultrasound data straight from a transducer that is
capable of acquiring volume RF (native) data in a standard way. This
is significantly richer in information than the raw or grey-level data that
are displayed by ultrasound machines, and is not affected by any of the
machine’s user settings. Statistical and mathematical analysis of all of
the data captured from such an ultrasound transducer makes it possible
for HistoScanning to identify differentiated tissue and assess the size and
accurately locate the site of the differentiated tissue. Thus, potentially,
changes in specific lesions between one follow-up and the next can be
quantified. Figure 1 illustrates this process.
Seeing Is Believing – The Need for Clinical Evidence
As discussed earlier, all existing diagnostic imaging technologies appear
to lack at least some of the key characteristics required to meet current
or anticipated clinical diagnostic needs. Furthermore, it is difficult to draw
direct comparisons between the different technologies in terms of
sensitivity and specificity due to the methodological differences in the
published studies (see Table 211,22,23). However, from the review of
constraints imposed on them in order that they can be visualised on a
the literature presented by Mark Emberton, it would appear that only
computer display, an additional two orders of magnitude of detail have
dynamic contrast-enhanced MRI (DCE-MRI) comes close to meeting the
been irretrievably ‘lost’ in the process. New technology such as
standards required clinically, but it still falls short of an ideal sensitivity or
HistoScanning may harness this otherwise ‘unused data’ to enhance the
specificity, even for lesions ≥0.5cc.
EUROPEAN UROLOGICAL REVIEW
17
Fitzpatrick_edit.qxp
25/9/08
10:50 am
Page 18
Prostate Cancer
Table 2: Sensitivity and Specificity of Different Imaging
Technologies for Detecting Prostate Cancer
Outcomes
Technology Dynamic
Contrast-enhanced
MRI22
Study
Prostate cancer
Population patients
Reference Radical prostatectomy
Method
histology
Tumour
>0.2cc >0.5cc
Volume
Sens. 77%
90%
Spec. 91%
88%
Acc. –
–
PPV 86%
77%
NPV 85%
95%
Figure 2: HistoScanning Visualisation of Extraprostatic Extension
Realtime
Elastography11
Power Doppler Contrast-enhanced
Ultrasound11
ADF Doppler 23
Raised PSA
Raised PSA
Raised PSA
Systematic biopsy Systematic biopsy Systematic biopsy
(transperineal)
(transperineal)
(transrectal)
–
–
–
68%
81%
76%
–
–
70%
75%
73%
–
–
100%
48%
–
–
–
required
MRI = magnetic resonance imaging; ADF = advanced dynamic flow; PSA = prostate-specific
antigen; PPV = positive predictive value; NPV = negative predictive value; Sens. = sensitivity;
Spec. = specificity; Acc. = accuracy.
to
verify
HistoScanning
against
step-sectioned
histopathology. Unconventional sagittal step-sectioning of the
prostate allowed for more accurate correlation to the sagittal
ultrasound scan, but made it difficult to sample the lateral margins of
Before it can be used routinely, the reliability and clinical value of any
the gland. As a result of the sagittal sectioning, the urethra was not
new diagnostic method must be firmly established. The first step
available as an ‘anchor’ for the registration of the HistoScanning
towards obtaining the level of evidence needed to establish prostate
against the histological grids. This may have resulted in some
HistoScanning’s potential was to explore and define its accuracy in
registration inaccuracies. There are no clear criteria for what
locating and sizing prostate lesions compared with the most reliable
constitutes one lesion. When is differentiated tissue one lesion, and
reference test, step-sectioned histology.24,25 The population studied
when is it two or more? There appears to be no clear consensus on
consisted of 29 men diagnosed with prostate cancer that was thought
which is the most appropriate measure for comparing the index and
to be confined to the prostate and was clinically attributed to T1c-
reference tests. There is no clear guide concerning total tumour
stage cancer and who were scheduled for radical prostatectomy. All
volume, number and volume of index and satellite lesions or even
patients were subjected to HistoScanning prior to surgery. Data from
when cancer is clinically ‘significant’.
15 of the patients were used to refine the HistoScanning analysis
algorithms, and HistoScanning analysis was performed ‘blind’ on the
Future Studies
remaining 14 patients without knowledge of the histology results.
Further multicentre studies in larger pertinent study populations are
(Data from one of the 14 patients was excluded from the final analysis
being implemented to both confirm the findings of the initial studies
due to damaged histology samples.)
and further explore HistoScanning’s capabilities and performance
under different clinical conditions. In order to establish the true
The results showed that prostate HistoScanning was able to accurately:
sensitivity and specificity of the test, the ideal study population consists
of men presenting with high PSA in whom no biopsy has been
• distinguish all cancer lesions ≥0.5cc from background tissue;
performed. However, the challenge is to find an accurate reference
• determine lesion location (100% concordance in determining focality
test approaching the sensitivity and specificity of step-sectioned
and laterality of the lesions); and
• estimate lesion size – prostate HistoScanning predictions correlate
closely with tumour volume as determined by planimetry at histology
histology. As standard biopsies sample less than 5% of prostate tissue,
it is unsuitable for this purpose. Therefore, template saturation
biopsies need to be considered as the reference test.
(r=0.97; p<0.0001).
All future studies will apply HistoScanning to the unprocessed RF volume
Furthermore, there were indications that prostate HistoScanning may
source data rather than the processed greyscale ultrasound data. It is
also be helpful for staging. For example, extraprostatic extension (EPE)
hoped that the vastly larger volume of data obtained this way will allow
was predicted in all cases that were later confirmed by histology (see
prostate HistoScanning to assess tumour aggressiveness, a major
Figure 2).
determinant of whether a tumour focus is considered clinically relevant.
Limitations of the Published Studies
Participant Perspectives on the Potential Clinical Value of
The first studies had certain methodological shortcomings, most of
Prostate HistoScanning
which will be addressed in ongoing and future studies. HistoScanning
After the above presentations, the discussion was opened up to the
analysis was based on standardised greyscale ultrasound volume files
colloquium participants and the faculty. Professor Fitzpatrick
rather than the unprocessed RF volume files, which contain two orders
encouraged the participants to share their thoughts about clinically
of magnitude more data. The study population was relatively small and
relevant potential uses, and also raise their questions and concerns
from a single centre (where a population screening programme was
about the HistoScanning technology. The areas where the participants
yet to be implemented). Patients were all known to have prostate
perceived HistoScanning to potentially offer the most benefit are
cancer and were scheduled for radical prostatectomy. This was
outlined below.
18
EUROPEAN UROLOGICAL REVIEW
Fitzpatrick_edit.qxp
25/9/08
10:50 am
Page 19
HistoScanning™ and Its Role in Prostate Cancer Diagnosis, Staging, Treatment and Monitoring
Diagnosis
Figure 3: Anterior Lesion Visualised by HistoScanning
If used as a triage test, HistoScanning may allow some patients
presenting with high PSA to avoid biopsy.
Targeting Biopsies
The possibility of directing biopsies to identified target areas was
thought to be of particular interest and created a lot of excitement. It
was felt that this application could substantially change the way in
which prostate cancer is diagnosed.
Detection of Anterior Tumours
It is difficult to biopsy in the anterior prostate, so a method of
visualising tumours in that sector would be particularly helpful in
clarifying the status of some of the high-PSA/negative-biopsy patients.
Since tumour location does not seem to influence HistoScanning’s
analysis, the technology is also promising for detecting anterior
tumours (see Figure 3).
effectiveness, it is important to know how the treatment or the
inherent tissue quality may affect HistoScanning results. The following
Staging
conditions were brought up by the participants and discussed.
HistoScanning may be able to provide an alternative to MRI or
supplement it as a technique for staging prostate cancer when MRI is
Calcifications
not readily available. Planned future studies will examine how reliably
Large calcifications cause ultrasound ‘shadowing’, or voids in
HistoScanning can indicate whether a lesion is confined to the prostate.
ultrasound data, and as such will affect the quality of the data
captured and available for HistoScanning. Calcifications can also affect
Targeting Treatment
step-sectioning of the prostate, which affects verification using the
Significant advantages were also seen in being able to concentrate a
reference method.
treatment on the tumour area. For example, in intensity-modulated
radiation therapy (IMRT), treatment could be concentrated on the
Chronic Prostatitis
tumour and the dose reduced elsewhere. In focal therapy, just the area
Based on available clinical data it is believed that chronic prostatitis will
of the tumour and its margins could be treated, followed by
not affect HistoScanning. This, along with all other possible
verification that the area had been effectively treated. In
pathologies, will continue to be evaluated in future studies.
brachytherapy, the dosimetry could be based on knowledge of the
Impact of Previous Therapeutic Interventions
tumour position.
The impact of treatments such as brachytherapy, high-intensity
Treatment Planning – Proximity to Neurovascular Bundle
focused ultrasound (HIFU), cryotherapy, radiotherapy, previous
Another potential use for prostate HistoScanning may be to see how
biopsies and hormone treatment has not yet been the subject of
close a tumour is to the neurovascular bundle.
clinical investigation. The impact of ultrasound shadowing by the
radioactive seeds and tissue destruction on the usefulness of
Monitoring
HistoScanning after brachytherapy must be investigated. It is believed
The ability to accurately monitor tissue changes over time facilitates active
that radiotherapy, HIFU and cryotherapy will dramatically change the
surveillance programmes as well as monitoring responsiveness to therapy.
nature of the tissue and the way it reflects ultrasound waves. Such
changes will probably be significant enough for treated tissue to be
Screening
differentiated from living tissue. This will no doubt be the subject of
Ultimately, prostate HistoScanning could find use as a screening
future studies.
method, replacing PSA testing.
Summary
Potential Limitations
The consensus among the participants was that the HistoScanning
As with any diagnostic technology, it is acknowledged that specific
results presented were exciting, and that if prostate HistoScanning lives
conditions or treatment effects may limit the effectiveness of
up to its apparent potential, the new technology may radically alter the
HistoScanning. If the tool is used for monitoring treatment
way in which prostate cancer is diagnosed and treated. ■
1.
2.
3.
4.
5.
6.
7.
8.
National Institute for Health and Clinical Excellence (NICE) clinical
guideline 58, Prostate Cancer: diagnosis and treatment, 2008.
Albertsen P, Cancer, 2008;112:2664–70.
Ahmed HU, et al., Nature Clin Prac Oncology, 2007;4(11):632.
Ikeda T, Shinohara K, Int J Urol, 2008;15:190.
Ellis WJ, et al., J Urol, 1994;152:1520.
Yu KK, Hricak H, Radiol Clin North Am, 2000;38:59.
Resnick MI, et al., J Urol, 1997;158:856.
Shigeno K, et al., BJU Int, 2003;91:223.
EUROPEAN UROLOGICAL REVIEW
9.
10.
11.
12.
13.
14.
15.
16.
17.
Bogers HA, et al., Urology, 1999;54:97.
Mitterberger M, et al., Eur Urol, 2008;53:112.
Kamoi K, et al., Ultrasound Med Biol, 2008; in press.
Masakazu T, et al., Urology, 2005;66:3.
Hricak H, et al., Radiology, 1987;162:331.
Coakley FV, et al., Radiology, 2002;223:91.
Yu KK, et al., Radiology, 1997;202:697.
DiBiase SJ, et al., Int J Radiat Oncol Biol Phys, 2002;52:429.
Engelbrecht MR, et al., Radiol, 2003;229:248.
18.
19.
20.
21.
22.
23.
24.
25.
Shukla-Dave A, et al., Radiology, 2004;231:717–24.
Haseman MK, et al., Clin Nucl Med, 1996;21:704.
Herrmann K, et al., J Nucl Med, 2004;45:359.
Resnick M, et al., Health Publications, 2006;186.
Villers et al., J Urol, 2006;176:2432–7.
Taymoorian K, et al., Anti Cancer Res, 2007;27(6C):4315–20.
Braeckman J, et al., BJU Int, 2007;101:293.
Braeckman, et al., BJU Int, 2008; in press.
19
`