The Limitations of Frozen Section?


 In the seventh Addition of Ackerman's Surgical Pathology, Juan Rosai (1) the frozen section is described as "one of the most important and difficult procedures a pathologist performs during his practice. It requires experience, knowledge of clinical medicine and pathology, the capacity to make quick decisions under pressure, good judgment, an attitude that is conservative but not excessively so, and a keen awareness of the limitations of the method". Well, what are these limitations? The Manual of Surgical pathology (2) in a section titled" Frozen Sections Are Not Permanent Sections" points to four reasons. These are: sampling error, ice crystal artifacts, lack of special studies, and lack of consultation.

          I would like to discuss what I consider the limitations in frozen section. I will divide these into two categories. True limitations I consider insurmountable limitations by virtue of the technique and constraints placed on it by the urgency of the procedure. Avoidable limitations I consider handicaps that we are forced to deal with but can be minimized to a degree where our ability to provide the diagnostic information can potentially match that of permanent sections.

True Limitations:

1) Time - There is no question that in the frozen section room there is often pressure to quickly come up with an answer. In my experience the best ways to make a mistake are to be rushed or interrupted. Our best defense is confidence, and good rapport  and education of our surgical colleagues. When pushed to speed up a frozen section we should resist this pressure at all costs. If you are a well trained and skilled frozen sectionist, the time to cut and stain a slide will be only a few minutes. If you are the pathologist grossing the tissue this process can take from a few minutes to 10 minutes or more for a large complicated specimen requiring multicolor inking. For the pathologist reading the slide, this can vary from seconds to 10 or more minutes if you are searching for some minute clue, paging through books or consulting colleagues. We must be practical and consider our surgical colleagues, but on the other hand, if this frozen section was requested for the proper reason we are asked to make a decision that will alter the course of surgery. We owe it to the patient to try our best to provide a correct answer, even if it is delaying the case a few extra minutes. It is far less costly than a re-operation or providing a wrong answer. If we are swamped with a barrage of cases, complicated cases or numerous specimens, the best we can do is ask for help. If any of the multitudes of specimens is not going to influence a surgical decision, then these should be put to the end of the line or not done at all. Try not to cut corners on the grossing, preparation and reading of the slides. This is where your errors will be born. One thing experience teaches you is to recognize a situation beyond your ability. In many of the cases in which I seek outside consultation, I know I'm stumped in the first few minutes of seeing the slide. It's like seeing an animal in the zoo for the first time. You're pretty sure you have never seen that in your backyard before. When faced with these cases I ask for whatever help is near. My advice is to tell your surgical colleagues all that you can be certain of and let them know they will have to be patient.

2) Limited special stains and studies - This is hard to argue with. Lets face it, in this day and age it would be malpractice try and sub classify lymphomas, sarcomas and all the things that mimic them without a few hardy trays full of immunoperoxidase stains and inspection of every twisted gene! Maybe the future will bring us more rapid studies that we can use at frozen section. As for now, this is a true limitation.

3) Lack of consultation - We are often alone in the frozen section room, with only our books to help us. Those of us in with large practices have the luxury of consultation with colleagues during working hours. The luxury of world expert consultation only an overnight express mail away is not an option. Or is it? At the present telepathology systems are being used to provide remote diagnostic services to distant hospitals. What began as a slow limited technology is now developing into an efficient practical means of outside consultation. I predict in the future that we will see telepathology develop into a widely used tool providing immediate intra-operative pathology consultation.

4)  Freezing artifacts -  It is a chemical property of water, that water will expand on freezing due to formation of hydrogen bonds. It is for this reason that ice floats. If it were not for this property the oceans would freeze over and we would not be here to think about it. Anyone who ever forgot a beer cooling in the freezer saw this principal in action.  I believe the changes we see in tissues which are frozen are related to this expansion of water upon freezing.  I will try to illustrate  the differences  between slides prepared by frozen section and those prepared by paraffin embedding.

Like any artifacts we deal with in pathology, recognizing the artifact allows us to "read around them" so that we can make the correct interpretation.

Below are phenomena which I believe are artifacts of freezing.

Ice Crystals ("bubbles") in  edematous stroma. Very edematous tissues freeze with an appearance similar to soap bubbles. I believe  as the water freezes the  expanding water forms rounded ice crystals which compresses the strands of fibrous tissue giving appearance of bubbles (Lower left) The frozen control shows that as the tissue is processed these "bubbles" of water redistribute into the stroma. The tissue in the right was never frozen and shows by comparison the edematous nature of the stroma.

Frozen Section

Frozen Control

Not Frozen

These very edematous tissues can  difficult to cut without shattering due to the icy consistency. These tissues  be as warm  as possible to get a clean section.

Compression Artifacts  Cellular tissues will be compressed by expanding ice bubbles. This is most evident in edematous tissues. The example of the kidney parenchyma below demonstrates  an extreme example of this phenomenon. The center picture shows the renal tubules being compressed by the clear ice crystals. The picture on the right shows the same tissue which was never frozen,

Frozen Section

Frozen Section

Not Frozen

Nuclear ice crystals   Nuclei will show varying tendency to form ice crystals. From my observations this seems to relate to the type of tissue as well as the state of the tissue. I have noticed more of these crystals in damaged tissues. It makes sense that tissue which are damaged by cautery or ischemia would have loss of osmotic homeostasis an might therefore result in more "nuclear edema". It also seems that the more vesicular nuclei have greater tendency to show these ice crystals. I have made one very important observation:

The thinner the tissue is cut the more these crystals appear as holes! The examples below clearly illustrate this point.



Lung Adenoarcinoma



Uterine Sarcoma

6 Microns

3 Microns

 As I mentioned earlier some times even if our cryostat is set for 6 microns we will get "thin sections" which actually are much thinner. This will explain why sometimes these crystals are more numerous.

Nuclear chromatin changes in frozen control. The example on the right below demonstrates the chromatin in this  previously frozen tissue. The chromatin is somewhat more condensed and hyperchromatic than the frozen section on the left which has been fixed rapidly. Notice the more vacuolated cytoplasm in the frozen sample, another very subtle example of freeze artifact.

Frozen Section

Permanent Section Previously Frozen

Permanent Section Never Frozen

Avoidable limitations:

1) Drying artifacts - A great deal of the nuclear artifact that I see in frozen sections is a result of "drying artifact." Just as the cells of an air dried diff quick stained slide have features of smudgy poorly define chromatin, the nuclei of a frozen section left for more than few seconds to dry show a loss of definition. A slide that can be fixed in a second or two in this time will show nuclear detail that rivals good cytology preparations. The exception to this will be the very vesicular nuclei which will retain its empty appearance which I consider a freezing artifact. (See above)

2) Sampling error One of my favorite questions to my residents is" What is the most important thing we do in the frozen section room?" The answer is "The gross". It does not matter how perfect our slide turns out if we have sampled the wrong part of the tissue. I insist that the process of grossing in the frozen section room should be performed a meticulous systematic fashion. If I handle a breast biopsy or other solid tumor all external aspects should be observed and palpated carefully before inking. Then after inking, slicing or dissection is performed. The specimen is laid out in an orderly fashion. Then starting from the first slice, the tissue is examined with the eyes and palpated with the fingers slice by slice. Complex organs are examined by anatomic regions. One must be careful not to jump to the obvious nodule in one of the central sections because by doing this surely one day they will miss the 3 mm nodule at the edge. Another piece of advice to my residents is "When your looking at a specimen if you cannot smell the tissue you are not looking close enough." If the surgeon is only supplying us with a small portion of tissue, than there will be sampling error (his), but this will apply equally to the permanent section. With a good gross examination the only sampling errors will be dictated by how many sections we have time to sample and cut. This is really the limitation of time and what's practical.

3) Fat - The yellow monster. This greasy guy just does not freeze. But there are approaches to deal with him that can help us through most situations. I was not sure which category to place fat. By virtue of the laws of chemistry and physic this is an inaqueous material which does not freeze. But it our real purpose of freezing tissue is to harden it so that it may be cut thin. And in fact at very low temperatures this material does harden. But at that temperature the tissues we are interested will shatter. It is unquestionably our supreme nemesis in the cutting of frozen sections and will always be a handicap. (See "Wrestling the fat one" on page FS technique III)

4) The quality of the section is inferior -  In the hands of an artist a frozen section can be cut and stained with the quality of a permanent section. If it's fixed quickly as I have described above the cytology can be preserved as well. Is the technique of frozen section to blame, or the inconsistent way it is taught? Can we blame a poorly stained slide on the stain or the person who stained it. Can we blame a wrinkles and shattered section on the technique or the technician? This brings us to the next avoidable limitation.

5) Inconsistency of training and performance - There is great deal of inconsistency in the training of frozen sectionists, be they histotechnologists, PA's or pathologists. There is inconsistency in who performs different parts of the task. Is the pathologist selecting the sections and dictating the approach to embedding? Is it a resident? A pathologist's assistant.? A histotechnologist? I mentioned above that I believe the gross to be extremely important. Even in a small sample, failing to recognize a minute gross detail could create an error due to poor embedding. A large complex specimen needs to be systematically examined by a pathologist without question. Once embedded, the frozen should be cut and stained by a properly trained individual. But again there is considerable variability in the training of these individuals.

6) Embedding - This is a great limitation to the conventional methods of preparation of frozen section. We may be asked to balance an icicle on the head of a pin. To accomplish this cryostats offer us a piece of steak on a tin plate and precede to squashes it with a hammer! Using conventional methods we may be asked to attempt to prepare precious minute tissues that may later disappear into a snowstorm. Up until now this is where the wizards had to resort to their magic. Through tedious and time consuming manipulations tissues are teased into various best attempts. The techniques I have offered in this web site will rapidly prepare blocks with a level of precision that surpasses paraffin embedding. I have resorted to using my limited artistic ability to demonstrate the level of facility and precision capable using these surprisingly simple techniques.



One must know what excellence looks like and sounds like in order to begin to approximate it. In all forms of art one can only achieve excellence if they are aware of all the ways to make mistakes and are aware when they are happening. SP

 In  this discussion I have attempted to provide a detailed fund of information which would offer operator a technical method to prepare high quality frozen sections. I have also tried to provide the details necessary to distinguish problems as they were arising and solutions with which to approach them. These are the techniques and observations I have gained in my experience. I have no doubt that there are many experienced pathologists and histotechnologists who have a multitude of valuable observations and techniques which remain unshared. It is my hope that others will be stimulated by these writings to present their own unique methods. My intention in these writings is to offer a nidus of information for colleagues to add to or criticize.

Note to readers: These techniques are mostly bases on my own observations in my practice. Anyone who has anything to add, suggestions to improve on this discussion or disagrees with any part of it please Email me at I will be happy to add useful information and credit the contributor. SP

1. Rosai, J. Chapter 1 Introduction, In: Ackerman's Surgical Pathology seventh Ed. Mosby, 1989:8-9.

2. Lester,SB. Chapter 6. Operating Room Consultations, In: Manual of Surgical Pathology. Churchill Livingston, 2001: 35-52