Dr Gonzalo E. Díaz Murillo


Diagnostic ultrasound
Interesting Answers and Questions

Question: Sonography and Statistics Accuracy

Answer: Ultrasonography accuracy is relative. Most sonograms result "normal" in spite of pathology. You can prove a condition exists but, you cannot prove it does not exist. It is very difficult to get accurate statistics if you cannot do any biopsy or autopsy to verify your results. There are not ultimate proves either.

The current technology would let us surpass operator dependence up to some important level but most manufacturers prefer to stick to known and proven technology rather than creating new things. The first is more profitable while the second involves expensive risks.

IMHO, the only way we can honestly protect our patients is by doing our best at all times while being conscious of the technology limitations and ours as well. More ...

Question: I am interested in finding out more information about the pros and cons of using US to diagnose solitary thyroid nodules, what other diagnostic modalities are available and why US may or may not be the modality of choice. Please forward to me your opinion or any other information you might have available.

Answer: Here are the pros of using ultrasonography to diagnose solitary thyroid nodules:

  1. It is safe (there is no risk at all) and you can scan the patients repeatedly and have proper follow-ups
  2. US can tell the difference between cysts and solid masses
  3. US can easily tell how close the masses are form important vessels
  4. US can detect changes consistent with malignancy
  5. US can easily measure dimensions and volume
  6. US is less expensive

Disadvantages: US cannot tell when a mass is hypo or hyperactive

Other diagnostic modalities you can use beside lab tests, cytology and biopsy are CT, Thermography, MRI and Gammagram.

Next 7 questions came from Fiona Soon Chin Fhong; Msc.in Medical Electronics University Technology of Malaysia.

Question: As a medical electronic student, i would like to seek for your stand point of view regarding the phenomenon called speckle noise. Does it has an effect during a scanning especially in early detection of pathologies? And, what do you think of the speckle effect during a scanning of obese person? Does the ultrasound machine offers a solution on that or can the sensitivity and contrast control reduce the problem?

Answer: Speckle noise is an undesirable artifact that masks pathology. It could be reduced using a phase insensitive technique, or a way to cancel the undesirable linear-phase representation. You could use a system sampling more images (more channels and memory needed). You could also consider adding a fuzzy logic algorithm. Be aware that if you do that, you are changing the image to meet your needs. That is the way ultrasound post-processing and pre-processing work though (changing the image to meet your needs).

The hardware requirements are difficult to meet now but if you use fuzzy logic algorithms, you could get outstanding results.

Speckle phenomena are common when scanning fat patients. Many scanners offer ways to somehow overcome this effect in a limited approach. One of my scanners - a digital one- uses special algorithms to cancel the speckle effect, but results are limited. Of course, digital imaging is a lot better. Gain and contrast controls offer poor improvement and pre and post-processing are better solutions (IMHO). The best practical solution now is scanning from different positions and angles and/or creating artificial acoustic windows.

Question: Through my recent reading in a paper by Dr. Aly A Farag ( IEEE Transaction in Medical imaging) , I was informed that the accuracy of using ultrasound diagnosis is about 78%. What about the other 28%? Do you agree with this statement?

Answer: I have not heard about it. It depends of the factors included in the survey. Sonography is a highly operator dependable procedure and results can vary much. If the statistics you mention consider different scanners, transducers, different operators, like sonologists, radiologists, sonographers, general physicians, care of the persons involved in the procedure (Private practice and research sonograms are more accurate), etc., the 78% would be accurate. No system is perfect.

In my personal experience, I have found ultrasonography being more accurate than CT and X-rays when dealing with abdominal and muscle pathology. Also found breast sonograms being a lot better than X-ray mammography and diagnostic laparoscopy. But I repeat: This is my experience. I have also seen absolutely wrong and useless sonograms. It is very relative. More ...

Question: Why are there exist minor cases of radiologist misdiagnosed or made inaccurate interpretation?

Answer: Minor - and mayor- cases of radiologists' misdiagnosis and inaccurate interpretations depend mainly on:

1- Insufficient knowledge.

2- Insufficient scanning time. Many services must scan a high number of patients a day to be profitable. I know of sites which pay for a pelvic sonogram less than USA $8.50. In the USA the same sonogram costs about USA $265.

3- Bad technique

4- Old and unreliable machines

Question: What are the difficulties found with the images by radiologist during an interpretation?

Answer: The most you talk with your patients, the better diagnosis you do. You could even get the diagnosis just by talking with the patient and practicing a limited physical examination. That is what I usually try to do, confirming my clinical diagnosis with my scanner but ...

Moved by technology, I set up a telesonography service over the Internet: I receive sonogram pictures images, some data relating patient symptoms and give back my diagnosis. I also set up a secondary and inexpensive ultrasound service, helped by a sonographer. I must be honest though: Both services are relatively cheap for the users because I cannot get as good results as doing the sonograms by my self. I will keep trying to get best and more economical results though, but also will keep my private practice scanning my patients.

Question: Is there something wrong with the machine resolving power or due to other reasons? I'm very curious about that.

Answer: Most of the time scanners work perfectly. Let me tell you something I saw years ago:

I was attending an ultrasound meeting. Manufactures came from many countries with their best machines. Important professors came from all over the world. They set up a live demonstration and presented a pregnant woman to be scanned in front of the audience. She told she was in her 18th week. A famous professor scanned the woman using a well known French machine. He -and us- saw nothing. Several slave monitors showed no pregnancy at all. The professor, using the microphone told: "This machine is not working well. It is useless. Please, could any one give a good machine?". A big "OOOOOOOOOOH" was heard coming from the astonished audience. The professor immediately got another equipment. Neither worked. He got another equipment, and other, and other, etc. None worked for the professor.

I realized the woman's bladder was overfilled and displacing the pregnant uterus upwards out of the pelvis. I left my seat and went behind the scenes, and told them what was happening. Incidentally, the woman was the wife of a friend of mine and I could verify that she really was in her 18 week.. The professors also verified it .... behind the scenes . . .

Note a single word was told to the audience.

Question: During my last visit to Batu Pahat General Hospital in Malaysia, I was told by the radiologist that what is to be solved is to resolve the parenchymal changes on tissue and if possible find a solution in early detection of Hepatitis B. Do you agree with him?

Answer: You can resolve parenchymal changes that could be consistent with hepatitis, but you cannot assure it is hepatitis.... yet

Question: Last question, how much does a radiologist depend on the clinical test report during the scanning?

Answer: A lot. Non only on the scan report, but on the questions and answers the radiologist and patients exchange and on the limited physical examination the radiologist should perform before scanning

Question: How accurate are obstetric sonograms?

Answer: It depends mainly on these factors:

  1. Gestational age: Individual differences increase as pregnancy progresses so, measurements taken early in pregnancy are most accurate
  2. Used measurement: CRL measurement is more accurate than gestational sac volume, gestational sac volume is more accurate than measure just one GS diameter, etc.
  3. Operator skill and experience: To get precise measurements, you must place the calipers according to proper anatomic landmarks
  4. Scanning time: Faster examinations are less accurate that dedicated examinations
  5. Hardware: Intracavitary probes provide higher frequency, so their resolution is better. Digital machines also do it better than analog machines
  6. Software: Its precision depends of statistics accuracy and on the software algorithm. Dedicated external software is more accurate that the software the ultrasound scanners provide

Next statistics come from my own experience, when doing my own sonograms, considering that I have 19+ year sonography experience, dedicate necessary time for each examination (just private practice), use a very good digital machine with probes from 2.5 MHZ up to 8.5 MHz , and dedicated and updated software developed by myself, out of statistics obtained with appropriate feedback. Although they are very accurate, keep in mind that the decimal factor result from calculations and they does not mean accuracy. For example, 4.1 days, does not mean that accuracy goes as high as 0.1 day (2 hours. 24 minutes). They are displayed because they are used for calculations precision. More ...

Weeks Error (days)

3.9 ---> 3.0
6.1 ---> 2.6
7.2 ---> 3.0
9.8 ---> 4.1
13.1 ---> 5.5
15.0 ---> 6.3
19.3 ---> 10.0
27.4 ---> 14.2
39.5 ---> 20.5

I have been reluctant to use my own software when providing obstetric telesonography service because conditions are different and I could not rely on my statistics